1.Application value of machine learning prediction model for neural invasion in gallbladder cancer based on enhanced CT and clinical characteristics
Bing ZHOU ; Sheng ZHANG ; Hao LI ; Binjie ZHOU ; Yang JIAO ; Qingwu WU ; Junyan YUE ; Shaoying LI
Chinese Journal of Digestive Surgery 2025;24(4):535-542
Objective:To explore the application value of machine learning prediction model for neural invasion in gallbladder cancer based on enhanced computed tomography (CT) and clinical characteristics.Methods:The retrospective cohort study was conducted. The clinical and imaging data of 502 patients with gallbladder cancer who were admitted to The First Affiliated Hospital of Xinxiang Medical University from January 2010 to June 2024 were collected. There were 171 males and 331 females, aged 65(range, 35?91)years. All patients underwent preoperative abdominal enhanced CT and radical resection. The 502 patients were randomly divided into a training set of 351 cases and a test set of 151 cases at a 7:3 ratio. The training set was used to construct prediction model, and the test set was used to validate prediction model. Observation indicators: (1)neural invasion in gallbladder cancer and influencing factor analysis; (2) construction and validation of machine learning prediction models for neural invasion in gallbladder cancer. Comparison of count data between groups was conducted using the chi-square test. Comparison of ordinal data between groups was conducted using the Mann-Whitney U test. Logistic regression model was performed for univariate and multivariate analyses. Independent influencing factors were incor-porated to construct machine learning models using the standard library modules based on Python 3.9. Receiver operating characteristic (ROC) curves were plotted, and the accuracy, sensitivity, specificity, area under the curve (AUC), precision, F1 score, positive predictive value, negative predic-tive value, and Kappa value were calculated to evaluate the predictive performance of the models. The Delong test was used to assess the differences in AUC among different models in the test set. The Hosmer-Lemeshow test and Brier score were used to evaluate the calibration of the models. Results:(1) Neural invasion in gallbladder cancer and influencing factor analysis. Of the 502 patients with gallbladder cancer, 131 cases had neural invasion, and 371 cases had no neural invasion. Results of multivariate analysis showed that total bilirubin, carcinoembryonic antigen, CA199, CA125, neutrophil-lymphocyte ratio, liver invasion detected by CT, vascular invasion detected by CT, hilar or retroperi-toneal lymph node metastasis detected by CT, and tumor stages T3 and T4 were independent influencing factors for neural invasion in patients with gallbladder cancer [ odds ratios=3.747, 2.395, 3.917, 3.596, 2.805, 2.377, 3.523, 2.774, 5.080, 6.809, 95% confidence interval ( CI) as 1.890?7.430, 1.154?4.971, 2.054?7.472, 1.807?7.155, 1.506?5.225, 1.241?4.553, 1.666?7.449, 1.483?5.189, 2.050?12.589, 2.552?18.168, P<0.05]. (2) Construction and validation of machine learning predic-tion models for neural invasion in gallbladder cancer. Based on the independent influencing factors, seven machine learning models were constructed, including logistic regression, K-nearest neighbors, support vector machine, random forest, decision tree, back-propagation neural network, and gradient boosting machine. The ROC curves of seven machine learning models in the test set were plotted, and the AUC were 0.900(95% CI as 0.851?0.948), 0.741(95% CI as 0.646?0.829), 0.836(95% CI as 0.762?0.895), 0.782(95% CI as 0.701?0.855), 0.839(95% CI as 0.770?0.901), 0.817(95% CI as 0.738?0.887), 0.843(95% CI as 0.770?0.909), respectively. Results of Delong test showed that the logistic regression model had the highest AUC. The sensitivity and specificity of the logistic regression model were 0.868 and 0.805 respectively, indicating the best balance. Results of Hosmer-Lemeshow test showed that the logistic regression model had a good goodness-of-fit ( χ2=5.320, P>0.05). The Brier score of the logistic regression model was relatively low, as 0.168, which verified its calibration advantage. Conclusion:Total bilirubin, carcinoembryonic antigen, CA199, CA125, neutrophil-to-lymphocyte ratio, liver invasion detected by enhanced CT, vascular invasion detected by enhanced CT, hilar or retroperitoneal lymph node metastasis detected by enhanced CT, and tumor stages T3 and T4 are independent influencing factors for nerve invasion in patients with gallbladder cancer. Seven machine learning models are constructed based on enhanced CT and clinical characteristics to predict neural invasion in gallbladder cancer, of which the logistic regression model demonstrates good predictive performance.
2.Application value of Bama miniature pig model in endoscopic retrograde cholangiopancrea-tography training
Sheng CHEN ; Ping2 YUE ; Yanyan2 LIN ; Jinduo2 ZHANG ; Long2 GAO ; Ling'en3 ZHANG ; Zhengping AN ; Chenyang WEI ; Pengfei LI ; Xun2 LI ; Wenbo2 MENG
Chinese Journal of Digestive Surgery 2025;24(7):898-904
Objective:To investigate the application value of Bama miniature pig model in endoscopic retrograde cholangiopancreatography (ERCP) training.Methods:The experimental study was conducted. Six Bama miniature pigs were selected to establish animal model for ERCP operation. The experimental equipments were the EP-6000 image processing system and ED-580T therapeutic duodenoscopy, and the surgery was performed by two experienced chief physicians. Observation indicators: (1) surgical operation status; (2) preoperative and postoperative physiological and biochemical indicators of experimental animals; (3) experimental endpoint indicators. Measurement data with normal distribution were represented as Mean± SD, and measurement data with skewed distribution were represented as M(range). Count data were expressed as absolute numbers. Results:(1) Surgical operation status. During the operation, the time required for the duodenoscope to pass from the esophageal inlet to the pylorus of the six experimental animals was 151(range, 46-227)seconds, the insertion length of the endoscope at the pylorus was 85(range, 80-90)cm, and the straightened length of the endoscope shaft was 47(range, 45-50)cm. For bile duct cannulation, the total number of cannulation attempts was 2.0(range, 1.0-3.0), the number of successful cannulation was 1.5(range, 1.0-2.0), and the time of successful cannulation was 44(range, 13-80)seconds. For pancreatic duct cannulation, the total number of cannulation attempts was 3.0(range, 1.0-8.0), and only one successful intubation was performed for each experimental animal, with the time of successful cannulation as 100(range, 53-320)seconds. (2) Preoperative and postoperative physiological and biochemical indicators of experimental animals. For six experimental animals, the preoperative body temperature was (38.8±0.1)℃, serum amylase was (55.3±0.2)U/dL, serum lipase was 232.9(range, 29.4-547.3)U/L, alanine aminotransferase was (51±10)U/L, aspartate aminotransferase was (38±12)U/L, and total bilirubin was (3.2±0.7)μmol/L. At 24 hours postoperatively, the body temperature was (38.6±0.1)℃, serum amylase was (53.0±6.0)U/dL, serum lipase was 62.4(range, 29.4-94.1)U/L, alanine amino-transferase was (54±5)U/L, aspartate aminotransferase was (67±27)U/L, and total bilirubin was (3.5±1.9)μmol/L. (3) Experimental endpoint indicators. At the experimental endpoint of 14 days after surgery, the bile duct length in the six experimental animals was (8.6±1.2)cm, and the bile duct inner diameter was (2.5±0.5)mm. The pancreatic duct length was 12.9(range, 3.7-20.8)cm, and the inner diameter was (1.6±0.2)mm. Results of histopathological examination revealed that both CD3 and CD68 staining were negative in the bile duct and surrounding hepatic tissue, as well as in the pancreatic duct and adjacent pancreatic tissue at the experimental endpoint of 14 days after surgery.Conclusion:Bama miniature pig model can be used for ERCP training.
3.Optimal b-Value Sets Based on Intravoxel Incoherent Motion in Pulmonary Solid Benign and Malignant Lesions
Wei WEI ; Heng LI ; Na ZHAO ; Chanjuan YU ; Xiuzheng YUE ; Zhiwei SHEN ; Xiangfei CHEN ; Sheng ZHANG ; Xiao YANG ; Yuedong HAN
Chinese Journal of Medical Imaging 2025;33(8):834-839
Purpose To quantitatively compare the diffusion parameters of mono-and biexponential diffusion-weighted imaging models,and to obtain optimal sets of b-values in diffusion-weighted MRI for obtaining monoexponential apparent diffusion coefficient(ADC)close to perfusion-insensitive intravoxel incoherent motion(IVIM)model ADC(ADCIVIM)in identifying of pulmonary solid benign and malignant lesions.Materials and Methods IVIM was performed in 40 patients with solid nodule and masse in Xi'an Gaoxin Hospital from July 2021 to August 2022 using a 3.0T MR imager.Two experienced diagnostic radiologists subjectively evaluated the IVIM images.A single index model was used to calculate ADC values(ADC0-1 000,ADC20-1 000,ADC50-1 000,ADC80-1 000,ADC150-1 000,ADC300-1 000,ADC500-1 000,ADC300,500,1 000,ADC300,800,1 000,ADC300,500,ADC300,800 and ADC300,1 000).The reference standard ADCIVIM value were calculated using a double-exponential model.The physician's measurements between two physicians were measured.The malignant and benign groups were compared and receiver operator characteristic curve for all parameters were analyzed.Results The measurement consistency of ADC values under b value sets and ADCIVIM was very good,and the intraclass correlation coefficient was more significant than 0.75.The differences between ADCIVIM and ADC values in each b group were statistically significant(t=-6.016--2.500,all P<0.05).The area under the curve(AUC)of ADCIVIM was the largest(0.906),with an optimal threshold of 1.271×10-3 mm2/s,a sensitivity of 80.0%and a specificity of 93.0%.The diagnostic efficacy close to ADCIVIM were ADC300,800(AUC=0.891),ADC50-1 000(AUC=0.827)and ADC300,800,1 000(AUC=0.795),respectively.The optimal threshold of ADC300,800 was 1.140×10-3 mm2/s,the sensitivity and specificity were 80.0%and 93.7%,respectively.Conclusion Combining b-values 300 s/mm2 and 800 s/mm2 is recommended as routine scanning parameters for identifying the insensitive monoexponential ADC between benign and malignant solid pulmonary lesions.
4.Study on mechanism of Jiawei Shaofu Zhuyu decoction in treatment of endometriosis fibrosis based on mitophagy
Can-can HUANG ; Wen-wen WAN ; Xiu-jia JI ; Bin YUE ; Yu-gui ZHANG ; Xiao-hua ZHANG ; Li LIANG ; Guo-lian CHEN ; Quan-sheng WU ; Hai-yan MAO
Chinese Pharmacological Bulletin 2025;41(6):1177-1185
Aim To explore the mechanism of Jiawei Shaofu Zhuyu decoction in antagonizing endometriosis fibrosis by regulating mitophagy.Methods After the animal model was constructed,the syndrome was evalu-ated by general condition,organ water content and ther-mal imaging.The curative effect was evaluated by the weight of ectopic focus and the degree of adhesion.The pathological changes were compared using HE stai-ning,transmission electron microscopy,Masson and Sir-ius red staining.The expression of PINK1 and Parkin was detected by immunohistochemistry.The expression of mRNA and protein was determined by qPCR and Western blot,and the level of serum ROS was detected by ELISA.Results The autonomic activity of model mice was weakened,the water content of organs rose,and the temperature of limbs and lower abdomen was reduced by thermal imaging.HE staining showed obvi-ous hyperplasia of ectopic epithelium and glands.Transmission electron microscopy showed mitochondrial and endoplasmic reticulum structure damage,and nor-mal autophagy structure disappeared.Masson and Siri-us red staining showed increased collagen deposition;immunohistochemistry showed decreased expression of PINK1 and Parkin in ectopic foci.qPCR and Western blot showed that the expression of PINK1,Parkin,Bec-lin1,LC3 mRNA and protein in ectopic foci of model mice decreased,the expression of p62 mRNA and pro-tein increased,and serum ROS increased.The syn-drome performance of model mice was improved after the intervention of Jiawei Shaofu Zhuyu decoction;the inflammatory infiltration of ectopic foci was relieved,the morphology of mitochondria and endoplasmic retic-ulum was restored,and normal autophagy structure ap-peared.The degree of collagen deposition and fibrosis was reduced;the mRNA and protein expression of PINK1,Parkin,Beclin1 and LC3 increased.The ex-pression of p62 mRNA and protein decreased,and the level of ROS decreased.Conclusions Jiawei Shaofu Zhuyu decoction can improve the fibrosis of ectopic le-sions in mice with endometriosis of cold-dampness sta-sis syndrome,which may be related to the regulation of mitophagy.
5.Study on mechanism of Jiawei Shaofu Zhuyu decoction in treatment of endometriosis fibrosis based on mitophagy
Can-can HUANG ; Wen-wen WAN ; Xiu-jia JI ; Bin YUE ; Yu-gui ZHANG ; Xiao-hua ZHANG ; Li LIANG ; Guo-lian CHEN ; Quan-sheng WU ; Hai-yan MAO
Chinese Pharmacological Bulletin 2025;41(6):1177-1185
Aim To explore the mechanism of Jiawei Shaofu Zhuyu decoction in antagonizing endometriosis fibrosis by regulating mitophagy.Methods After the animal model was constructed,the syndrome was evalu-ated by general condition,organ water content and ther-mal imaging.The curative effect was evaluated by the weight of ectopic focus and the degree of adhesion.The pathological changes were compared using HE stai-ning,transmission electron microscopy,Masson and Sir-ius red staining.The expression of PINK1 and Parkin was detected by immunohistochemistry.The expression of mRNA and protein was determined by qPCR and Western blot,and the level of serum ROS was detected by ELISA.Results The autonomic activity of model mice was weakened,the water content of organs rose,and the temperature of limbs and lower abdomen was reduced by thermal imaging.HE staining showed obvi-ous hyperplasia of ectopic epithelium and glands.Transmission electron microscopy showed mitochondrial and endoplasmic reticulum structure damage,and nor-mal autophagy structure disappeared.Masson and Siri-us red staining showed increased collagen deposition;immunohistochemistry showed decreased expression of PINK1 and Parkin in ectopic foci.qPCR and Western blot showed that the expression of PINK1,Parkin,Bec-lin1,LC3 mRNA and protein in ectopic foci of model mice decreased,the expression of p62 mRNA and pro-tein increased,and serum ROS increased.The syn-drome performance of model mice was improved after the intervention of Jiawei Shaofu Zhuyu decoction;the inflammatory infiltration of ectopic foci was relieved,the morphology of mitochondria and endoplasmic retic-ulum was restored,and normal autophagy structure ap-peared.The degree of collagen deposition and fibrosis was reduced;the mRNA and protein expression of PINK1,Parkin,Beclin1 and LC3 increased.The ex-pression of p62 mRNA and protein decreased,and the level of ROS decreased.Conclusions Jiawei Shaofu Zhuyu decoction can improve the fibrosis of ectopic le-sions in mice with endometriosis of cold-dampness sta-sis syndrome,which may be related to the regulation of mitophagy.
6.ALKBH5 mediated m6A modification of NLRP3 promotes cardiomyocytes pyroptosis in mice with myocardial infarction
Miao-miao ZHAI ; Jian-jian YIN ; Zhi-mo WANG ; Yue-jiao ZHOU ; Qing-wen YU ; Pei WANG ; Li-rong ZHANG ; Sheng-na HAN
Chinese Pharmacological Bulletin 2025;41(3):434-444
Aim To investigate the effects of m6A demethylase ALKBH5 on cardiomyocytes pyroptosis in mice with myocardial infarction(MI).Methods The MI model of left anterior descending coronary artery ligation surgery was established by knocking down ALKBH5 using adeno-associated virus,and the hypox-ia model of mouse cardiomyocytes(HL-1)was estab-lished by knocking down small interfering RNA.The effects of ALKBH5 on the pyroptosis of MI mice and hypoxic HL-1 cells were observed.Subsequently,mechanism studies were conducted at the cellular lev-el,and the binding of ALKBH5 and IGF2BP2 to NL-RP3 mRNA was detected through RNA pull down and RNA immunoprecipitation(RIP)experiments.The MeRIP-qPCR method was used to determine the effects of ALKBH5 on the mRNA m6A level of NLRP3.Acti-nomycin D for RNA stability experiments were conduc-ted to detect the effects of ALKBH5 and IGF2BP2 on the stability of NLRP3 mRNA.Results Knocking down ALKBH5 in vivo and in vitro both inhibited NL-RP3 inflammasome activation and alleviated pyroptosis in MI mice and hypoxic HL-1 cells.Mechanistically,the results showed that NLRP3 mRNA could bind to ALKBH5 protein in HL-1 cells;knocking down ALK-BH5 could increase the m6A level of NLRP3 and re-duce the stability of NLRP3 mRNA;subsequently,it was confirmed that NLRP3 mRNA and IGF2BP2 pro-tein bound to each other;knocking down IGF2BP2 in-creased the mRNA stability of NLRP3.The Rescue ex-periment showed that knocking down IGF2BP2 re-versed the decrease in NLRP3 mRNA expression caused by knocking down ALKBH5.Conclusions ALKBH5 mediated m6A modification of NLRP3 pro-motes cardiomyocytes pyroptosis in mice with myocardi-al infarction.
7.Advances in mechanism of mitochondrial quality control system in endometriosis
Yuan-huan CHEN ; Bin YUE ; Hai-yan MAO ; Can-can HUANG ; Xiao-hua ZHANG ; Peng FENG ; Quan-sheng WU
Chinese Pharmacological Bulletin 2025;41(3):406-410
Endometriosis(EMs)is a common estrogen-depend-ent clinical disease with the pathological characteristics of malig-nant tumors,which has great impact on women's physical and mental health.In recent years,experimental exploration has re-vealed that ectopic foci are in a hypoxic environment outside the uterine cavity,and mitochondria,as the"functional factories"of the cells,play an important role in the process of planting and in-vasion,and the mitochondrial quality control system,which in-cludes mitochondrial oxidative stress,kinetics,autophagy,bio-genesis and calcium homeostasis,is a key mechanism for the e-quilibrium of the mitochondrial function.The mitochondrial quality control system,including mitochondrial oxidative stress kinetics,autophagy,biogenesis and calcium homeostasis,is a key mechanism for mitochondrial functional balance.Therefore,to clarify the role of the mitochondrial quality control system in the development of EMs with the help of rational and rigorous experi-mental and clinical studies can not only help to clarify the patho-genesis of the disease,but also explore the key targets in the prevention and treatment of the disease.Therefore,this article summarizes the research progress of mitochondrial quality control system in endometriosis,with a view to providing reference and theoretical basis for the etiology,pathogenesis and prevention strategies of EMs.
8.The Improvement of Intestinal Damage and Lipid Metabolism in Diabetic Rats by the Regulation of IRAK1/NF-κB Signaling Pathway by"Srperficiality-Origin Matching Acupoints"Electroacupuncture
Congcheng WANG ; Xia SUN ; Yue SHENG ; Xiaoning ZHANG ; Fan PING
World Science and Technology-Modernization of Traditional Chinese Medicine 2025;27(1):211-219
Objective To investigate the improvement of interleukin-1 receptor associated kinase(IRAK1)/nuclear factor κB(NF-κB)signaling pathway regulated by"srperficiality-origin matching acupoints"electroacupuncture on intestinal damage in diabetic rats and lipid metabolism.Methods SD rats were randomly separated into control group,diabetes group,diabetes+electroacupuncture group,diabetes+pcDNA group,diabetes+pcDNA-IRAK1 group,diabetes+electroacupuncture+pcDNA group,and diabetes+electroacupuncture+pcDNA-IRAK1 group,18 rats in each group.Except for the control group,rats in other groups were fed with high glucose and high fat diet and intraperitoneal injection of streptozotocin(STZ)to construct the model of diabetes rats.After successful modeling,rats in each group were treated every other day for 8 consecutive weeks.The weight,fasting blood glucose,serum triglycerides and total cholesterol levels of rats were detected.16S rRNA amplicon sequencing was applied to detect the relative abundance of rat gut microbiota;the intestinal permeability and water content of rats were tested.Immunohistochemistry was applied to detect the average optical densities of zonula occludens-1(ZO-1)proteins in small intestine tissue.Western blot was applied to detect the expression of IRAK1,phosphorylated nuclear factor inhibitory protein(p-IκBα),and phosphorylated nuclear factor B p65(p-NF-κB p65)proteins in rat small intestine tissue.Results Compared with the control group,the levels of fasting blood glucose,triglyceride,total cholesterol,relative abundances of Firmicutes,Proteobacteria,Actinomycetes,Allobaculum,fluorescein isothiocyanate(FITC)-glucan concentration,intestinal water content and the expression of IRAK1,p-IκBα,p-NF-κB p65 proteins in the diabetes group increased significantly(P<0.05).Weight,the relative abundance of Bacteroidetes,Lactobacillus,Romboutsia,Lachnospiraceae,Ruminocoaceae,the average optical densities of occludin and ZO-1 proteins decreased significantly(P<0.05).Compared with the diabetes group,the levels of fasting blood glucose,triglyceride,total cholesterol,relative abundances of Firmicutes,Proteobacteria,Actinomycetes,Allobaculum,FITC-glucan concentration,intestinal water content and the expression of IRAK1,p-IκBα,p-NF-κB p65 proteins in the diabetes+electroacupuncture group decreased significantly(P<0.05).Weight,the relative abundance of Bacteroidetes,Lactobacillus,Romboutsia,Lachnospiraceae,Ruminocoaceae,the average optical densities of occludin and ZO-1 proteins increased significantly(P<0.05),the change trend of corresponding indicators in the diabetes+pcDNA-IRAK1 group was opposite to the above(P<0.05).Compared with the diabetes+electroacupuncture group and the diabetes+electroacupuncture+pcDNA group,the levels of fasting blood glucose,triglyceride,total cholesterol,relative abundances of Firmicutes,Proteobacteria,Actinomycetes,Allobaculum,FITC-glucan concentration,intestinal water content and the expression of IRAK1,p-IκBα,p-NF-κB p65 proteins in the diabetes+electroacupuncture+pcDNA-IRAK1 group increased significantly(P<0.05).Weight,the relative abundance of Bacteroidetes,Lactobacillus,Romboutsia,Lachnospiraceae,Ruminocoaceae,the average optical densities of occludin and ZO-1 proteins decreased significantly(P<0.05).Conclusion"Srperficiality-origin matching acupoints"electroacupuncture may improve intestinal flora imbalance and intestinal barrier damage in diabetes rats by inhibiting IRAK1/NF-κB signaling pathway,and improve diabetic glucose and lipid metabolism.
9.MiR-330-5p targets OY-TES-1 to inhibit the migration of glioblastoma
Guo LIANG ; Zhen-kai ZHAO ; Zhao-yue ZENG ; Qing-mei ZHANG ; Wei-xia NONG ; Xiao-xun XIE ; Xi-sheng LI
Journal of Regional Anatomy and Operative Surgery 2025;34(2):98-103
Objective To explore the targeted regulatory relationship of miR-330-5p on OY-TES-1 in glioblastoma and the effect of miR-330-5p/OY-TES-1 axis on the migration ability of glioblastoma.Methods Bioinformatics analysis was performed to analyze the expression level of miR-330-5p in patients with glioblastoma and its influence on prognosis and survival of patients.The glioblastoma cells U251 were divided into miR-330-5p minics group,minics-NC group,and miR-330-5p+OY-TES-1 overexpression group(miR-330-5p minics+pcDNA3.1-OY-TES-1).The effect of miR-330-5p on the activity of OY-TES-1 3'UTR region was detected by double luciferase reporter gene experiment.The expression of OY-TES-1 mRNA was detected by qRT-PCR.The effect of miR-330-5p/OY-TES-1 axis on the migration ability of glioblastoma cells was detected by Transwell migration assay.Results The expression of miR-330-5p in glioblastoma tissue was significantly lower than those in non-tumor brain tissue and low-grade glioma tissue(P<0.05).The survival time of glioblastoma patients with high expression of miR-330-5p was significantly longer than that of patients with low expression of miR-330-5p(P<0.05).After overexpression of miR-330-5p,the activity of OY-TES-1 3'UTR region was decreased(P<0.05).Compared with minics-NC group,the expression levels of OY-TES-1 mRNA of U251 and U87MG cells in miR-330-5p minics group were significantly decreased(P<0.01).Compared with minics-NC group,the numbers of migrating cells in miR-330-5p minics group and miR-330-5p+OY-TES-1 overexpression group were significantly decreased(P<0.05).Compared with miR-330-5p minics group,the number of migrating cells in miR-330-5p+OY-TES-1 overexpression group was significantly increased(P<0.01).Conclusion MiR-330-5p targets OY-TES-1 to inhibit the migration of glioblastoma.
10.Application value of machine learning prediction model for neural invasion in gallbladder cancer based on enhanced CT and clinical characteristics
Bing ZHOU ; Sheng ZHANG ; Hao LI ; Binjie ZHOU ; Yang JIAO ; Qingwu WU ; Junyan YUE ; Shaoying LI
Chinese Journal of Digestive Surgery 2025;24(4):535-542
Objective:To explore the application value of machine learning prediction model for neural invasion in gallbladder cancer based on enhanced computed tomography (CT) and clinical characteristics.Methods:The retrospective cohort study was conducted. The clinical and imaging data of 502 patients with gallbladder cancer who were admitted to The First Affiliated Hospital of Xinxiang Medical University from January 2010 to June 2024 were collected. There were 171 males and 331 females, aged 65(range, 35?91)years. All patients underwent preoperative abdominal enhanced CT and radical resection. The 502 patients were randomly divided into a training set of 351 cases and a test set of 151 cases at a 7:3 ratio. The training set was used to construct prediction model, and the test set was used to validate prediction model. Observation indicators: (1)neural invasion in gallbladder cancer and influencing factor analysis; (2) construction and validation of machine learning prediction models for neural invasion in gallbladder cancer. Comparison of count data between groups was conducted using the chi-square test. Comparison of ordinal data between groups was conducted using the Mann-Whitney U test. Logistic regression model was performed for univariate and multivariate analyses. Independent influencing factors were incor-porated to construct machine learning models using the standard library modules based on Python 3.9. Receiver operating characteristic (ROC) curves were plotted, and the accuracy, sensitivity, specificity, area under the curve (AUC), precision, F1 score, positive predictive value, negative predic-tive value, and Kappa value were calculated to evaluate the predictive performance of the models. The Delong test was used to assess the differences in AUC among different models in the test set. The Hosmer-Lemeshow test and Brier score were used to evaluate the calibration of the models. Results:(1) Neural invasion in gallbladder cancer and influencing factor analysis. Of the 502 patients with gallbladder cancer, 131 cases had neural invasion, and 371 cases had no neural invasion. Results of multivariate analysis showed that total bilirubin, carcinoembryonic antigen, CA199, CA125, neutrophil-lymphocyte ratio, liver invasion detected by CT, vascular invasion detected by CT, hilar or retroperi-toneal lymph node metastasis detected by CT, and tumor stages T3 and T4 were independent influencing factors for neural invasion in patients with gallbladder cancer [ odds ratios=3.747, 2.395, 3.917, 3.596, 2.805, 2.377, 3.523, 2.774, 5.080, 6.809, 95% confidence interval ( CI) as 1.890?7.430, 1.154?4.971, 2.054?7.472, 1.807?7.155, 1.506?5.225, 1.241?4.553, 1.666?7.449, 1.483?5.189, 2.050?12.589, 2.552?18.168, P<0.05]. (2) Construction and validation of machine learning predic-tion models for neural invasion in gallbladder cancer. Based on the independent influencing factors, seven machine learning models were constructed, including logistic regression, K-nearest neighbors, support vector machine, random forest, decision tree, back-propagation neural network, and gradient boosting machine. The ROC curves of seven machine learning models in the test set were plotted, and the AUC were 0.900(95% CI as 0.851?0.948), 0.741(95% CI as 0.646?0.829), 0.836(95% CI as 0.762?0.895), 0.782(95% CI as 0.701?0.855), 0.839(95% CI as 0.770?0.901), 0.817(95% CI as 0.738?0.887), 0.843(95% CI as 0.770?0.909), respectively. Results of Delong test showed that the logistic regression model had the highest AUC. The sensitivity and specificity of the logistic regression model were 0.868 and 0.805 respectively, indicating the best balance. Results of Hosmer-Lemeshow test showed that the logistic regression model had a good goodness-of-fit ( χ2=5.320, P>0.05). The Brier score of the logistic regression model was relatively low, as 0.168, which verified its calibration advantage. Conclusion:Total bilirubin, carcinoembryonic antigen, CA199, CA125, neutrophil-to-lymphocyte ratio, liver invasion detected by enhanced CT, vascular invasion detected by enhanced CT, hilar or retroperitoneal lymph node metastasis detected by enhanced CT, and tumor stages T3 and T4 are independent influencing factors for nerve invasion in patients with gallbladder cancer. Seven machine learning models are constructed based on enhanced CT and clinical characteristics to predict neural invasion in gallbladder cancer, of which the logistic regression model demonstrates good predictive performance.

Result Analysis
Print
Save
E-mail