自动乳腺全容积超声成像特征联合超声多参数评分在T1～T2期乳腺癌腋窝淋巴结转移预测中的增益价值OA

Objective To construct a prediction model based on automated breast volume scanner(ABVS)imaging features,ultrasound multi-parameter scoring and clinical factors,to analyze its predictive value for axillary lymph node metastasis(ALNM)in T1-T2 breast cancer,and to investigate the incremental value of ABVS imaging features and ultrasound multi-parameter scoring in the prediction model.Methods A total of 275 patients with T1-T2 breast cancer admitted from January 2017 to December 2024 were selected and divided into the metastasis group(n=103)and the non-metastasis group(n=172)according to the presence or absence of ALNM.The clinical data,ABVS imaging features,and ultrasound multi-parameter scoring were compared between the two groups.Multivariate logistic regression was used to analyze the influencing factors of ALNM in T1-T2 breast cancer.Clinical factor prediction model and imaging-clinical factor prediction model were constructed respectively,and visualized using nomograms.Receiver operating characteristic(ROC)and the area under ROC curve(AUC)were used to evaluate the predictive value of the two prediction models.Results The proportions of infiltrative type,T2 stage,and human epidermal growth factor receptor-2(HER-2)positivity in the metastasis group were higher than those in the non-metastasis group(P＜0.05).The maximum tumor diameter,proportions of convergence sign and calcification in the metastasis group were higher than those in the non-metastasis group(P＜0.05).The ultrasound elastography score,color Doppler ultrasound score,and ultrasound multi-parameter score in the metastasis group were higher than those in the non-metastasis group(P＜0.05).Pathological type,T stage,HER-2,maximum tumor diameter,convergence sign,calcification and ultrasound multi-parameter score were independent predictors of ALNM in T1-T2 breast cancer(P＜0.01).A clinical factor prediction model was constructed using pathological type,T stage and HER-2,and the imaging-clinical factor prediction model was constructed using pathological type,T stage,HER-2,maximum tumor diameter,convergence sign,calcification and ultrasound multi-parameter score,and the two prediction models showed good goodness-of-fit.The consistency index of the clinical factor prediction model and the imaging-clinical factor prediction model was 0.746 and 0.861 respectively,the calibration was 0.759 and 0.884,and the AUC was 0.707 and 0.860,respectively.The AUC of the imaging-clinical factor prediction model was significantly higher than that of the clinical factor prediction model(P＜0.05).Conclusion Pathological type,T stage,HER-2,maximum tumor diameter,convergence sign,calcification and ultrasound multi-parameter score are independent influencing factors of ALNM in T1-T2 breast cancer.The prediction model constructed based on the above factors can improve the prediction efficiency of ALNM in T1-T2 breast cancer.AB VS imaging features and ultrasound multi-parameter score have significant incremental value in the prediction model.