A risk prediction model of central lymph node metastasis nomograms based on parameters of thyroid contrast-enhanced ultrasound

QIU Shuang; GAO Jing; XIAO Shuang

doi:10.3969/j.issn.2097-1753.2025.04.012

Abstract：Objective To construct a risk prediction model for central lymph node metastasis(CLNM)in patients with papillary thyroid carcinoma(PTC)based on contrast-enhanced ultrasound(CEUS)parameters.Methods A total of 105 patients with PTC and treated in Tangshan People's Hospital between January 2022 and March 2023 were selected,and the incidence of CLNM was statistically analyzed.The patients were divided into the CLNM group and non-CLNM group.The basic information,indicators of laboratory tests,and CEUS parameters were compared between the two groups.The contributors to CLNM in patients with PTC were analyzed using a Lasso-Logistic regression model.The predictability and utility of the model were evaluated via the receiver operating characteristic curve and decision curve analysis.Results The incidence of CLNM in these PTC patients was 48.57%(51/105).The proportion of female patients,age,proportion of multiple lesions,maximum lesion diameter,scores of the thyroid imaging reporting and data system(TI-RADS),serum thymidine kinase 1(TK1)levels,gene expressions of targeting protein for Xklp2,C-X-C chemokine receptor 4(CXCR4),and those of a disintegrin and metalloproteinase 9(ADAM9)in cancer tissues in the CLNM group were higher than in the non-CLNM group(t/χ2=2.282～12.302,all P＜0.05).The serum levels of microRNA-363(miR-363)and microRNA-1296(miR-1296),gene expressions of microRNA-1243(miR-1243)and protein phosphatase 4 regulatory subunit 1,time to peak(TTP),peak intensity(PI),and mean transit time(MTT)were lower or shorter than in the non-CLNM group(t/χ2=5.914～8.638,all P＜0.001).The score of TI-RADS,TK1,miR-363,miR-1296,miR-1243,CXCR4,ADAM9,TTP,PI and MTT could make a difference in CLNM among PTC patients(OR=0.267～3.124,all P＜0.05).A risk prediction model of CLNM nomograms was constructed,and_Bootstrap internal validation showed that the area under the curve(AUC)of the model forpredicting CLNM in PTC patients was 0.915(95%CI:0.884～0.917);the decision curve(DCA)showed that the model had good clinical utility in predicting the risk of CLNM.The calibrationcurve showed that the coincidence between the observed value and the predicted value of thecalibration curve of the modcl was good.Conclusion The risk prediction model of CLNM nomograms based on CEUS parameters in PTC patients can well predict the risk of CLNM and promises wide clinical applications.