Three-dimensional shear-wave elastography for differentiating benign and malignant breast lesions: comparison with two-dimensional shear-wave elastography

Abstract

To evaluate the interobserver agreement and the diagnostic performance of 3D shear-wave elastography (SWE) for breast lesions in comparison with 2D SWE. A total of 163 breast lesions (malignant 48, benign 115) in 146 women who underwent B-mode ultrasound and SWE before biopsy were included. Two radiologists reviewed six data sets (B-mode, SWE, and a combination of both for 2D and 3D ultrasound). B-mode and SWE features were recorded. BI-RADS category was assigned for B-mode and combined sets. Interobserver variability was assessed using the κ statistic. Diagnostic performance of each data set was evaluated using the area under the ROC curve (AUC). SWE showed substantial to almost perfect agreement, with Ehomo in 2D SWE being higher than in 3D SWE. The AUC of 2D SWE was higher than 3D SWE for all SWE features, significantly so for Ecol (0.933 vs. 0.867, P = 0.002) and Emax (0.961 vs. 0.874, P = 0.006). After adding SWE to B-mode ultrasound, the AUC in 2D ultrasound increased significantly (0.968 vs. 0.912, P = 0.008), but 3D ultrasound showed no significant difference (0.966 vs. 0.935; P = 0.07). For 3D SWE, interobserver agreement was good, but the diagnostic performance was inferior to 2D SWE even after adding to B-mode ultrasound. • Shear-wave elastography (SWE) provides further diagnostic information during breast ultrasound. • 3D SWE diagnostic performance is inferior to 2D SWE. • In 3D SWE, interobserver agreement was good. • 2D B-mode ultrasound showed significant diagnostic improvement when combined with 2D SWE. • 3D B-mode ultrasound performance was not significantly improved when combined with 3D SWE.