Study on the predictive value of CT imaging features and clinical indicators in the pathological classification of pediat-ric hepatoblastoma

LI Qiancheng; WANG Wei; ZHANG Gongwei; LI Shixian; LI Chenglong; XU Chao; YU Xiao; WANG Han

doi:10.20258/j.cnki.1006-9011.2025.02.017

Abstract：Objective To analyze the differences of CT imaging features and clinical indicators of children with hepatoblas-toma(HB)and to explore the value of relevant indicators in the prediction of HB pathological classification.Methods 49 cases of HB admitted to Xuzhou Children's Hospital from January 2014 to December 2023 were included in this study,and di-vided into two groups according to pathological classification:mixed epithelial mesenchymal type(25 cases)and complete epi-thelial type(24 cases).The imaging and clinical data of all children with different pathological types were analyzed by Univariate analysis and Binary logistic regression,and the value of relevant indicators in HB pathological classification prediction was evalu-ated by the use of the receiver working characteristic curve(ROC).Results There were significant differences in age,calcifica-tion and envelope retraction among different pathological types.Binary logistic regression analysis showed that age and capsule re-traction were the independent influencing factors of HB in children.The area under the ROC curve of the evaluation model con-structed based on the above two indexes was 0.837(95%CI:0.720-0.953),the sensitivity and specificity were 81.76%and 95.47%respectively,which were higher than each single index in the model,and the fitting degree of the evaluation model was good(Hosmer-Lemeshow χ2=7.365,P=0.498).Conclusion Age and capsule regression are independent influencing factors to predict the pathological classification of HB in children.The evaluation model constructed based on the above two indexes has high predictive value for the pathological classification of HB in children,which can provide a basis for the clinical decision-making of HB in children.