Constructing a predictive model for benign and malignant breast lesions based on multimodal ultrasound imaging omics

WANG Xiaomei; LI Qiaoran; LI Xueli; ZHENG Hui; TAO Peng

Abstract：Objective To explore the construction of a predictive model for benign and malignant breast lesions based on multimodal ultrasound radiomics so as to achieve preoperative evaluation of breast lesions.Methods 78 patients with four types of breast mass lesions were selected using mammography and breast imaging report and data system(BI-RADS).The surgi-cal pathological results were used as the gold standard,and the patients were divided into malignant and benign groups,both of whom underwent multimodal ultrasound examination.CDFI and SWE parameters of patients were compared between the two groups.Results There were 52 cases of breast cancer in 78 patients with breast mass lesions,accounting for 66.67%(52/78).There were 26 cases of breast fibroadenoma,accounting for 33.33%(26/78).The CDFI examination results showed that the PSV and PI of the malignant group were higher than those of the benign group,and the difference was statistically significant(P<0.05).The SWE examination results showed that Emax and Emean in the malignant group were higher than those in the benign group,and the difference was statistically significant(P<0.05).The ROC curve was plotted,and the results showed that the AUC values of PSV,RI,Emax,and Emean for predicting benign and malignant breast mass type lesions were all≥0.70,both indi-vidually and in combination,and the predictive value of the combined examination was higher(Z=2.102,2.841,2.770,1.984,P=0.018,0.002,0.003,0.024).Conclusion A multimodal ultrasound imaging omics model based on PSV,RI,Emax,and Emean for breast mass type lesions can be used to distinguish between benign and malignant breast lesions.