Multimodal Ultrasound Characteristics of Small Breast Masses in Breast BI-RADS Category 4 to 5
Objective:Combined conventional ultrasound,ultrasound elastography and contrast-enhanced ultrasound were used to establish and validate apredictive model for the benignity and malignancy of small breast masses(≤2cm)in breast BI-RADS category 4 to 5.Methods:Ultrasound characteristics of 200 pathologically confirmed small breast masses were collected and subjected to Logistic regression analysis to establish amultimodal ultrasound diagnostic prediction model.A total of 59 patients with breast masses were also selected to validate the prediction model and plot ROC curves.Results:Multi-factor regression analysis showed that the last 7 features that entered the Logistic model were patient age(≥40years),shape andorientation of the mass,microcalcification,elastic strain ratio,increased lesion extent after enhancement,and local perfusion defect after enhancement.Conclusion:Multimodal ultrasound showed good diagnostic performance inpredicting benign and malignant lesions in breast BI-RADS category 4 to 5.
breast cancerstrain elastographycontrast-enhanced ultrasoundLogistic regression models
NETL
NSTL
万方数据
