Application of a shear-wave elastography prediction model to distinguish between benign and malignant breast lesions and the adjustment of ultrasound Breast Imaging Reporting and Data System classifications

Abstract

To explore a real-time shear-wave elastography (SWE) prediction model distinguishing benign from malignant breast lesions and to determine its application in adjusting ultrasound Breast Imaging Reporting and Data System (BI-RADS) classifications. Four hundred and sixty-eight patients with 488 breast lesions were enrolled. Patients underwent hollow-needle puncture or surgical resection for histopathological examinations. Ultrasound examinations, both conventional ultrasound and real-time SWE, were performed <2 weeks prior to sampling. Statistical analyses were implemented to distinguish benign from malignant breast lesions and adjust ultrasound BI-RADS 3 and 4a classifications. The real-time SWE indicators Emax and Ecol showed the highest diagnostic efficiency in distinguishing between benign and malignant lesions through quantitative and qualitative indicators, respectively. The area under the curve (AUC) for Emax was 0.837 while that for Ecol was 0.828. The AUC of the real-time SWE prediction model, constructed by multivariate logistic regression, for diagnosing benign and malignant breast lesions was 0.850. The real-time SWE prediction model aids in the differential diagnosis of benign and malignant breast lesions but cannot replace conventional ultrasound. The model improves the diagnostic performance of ultrasound BI-RADS 3 and 4a classifications.