AIM: The aim of this study is to retrospectively assess the additional diagnostic value of shear wave elastography (SWE) added to ultrasound (US) versus US alone in differentiating malignant and benign non-mass lesions (NMLs) of the breast by readers with different experience levels and to assess interreader agreement. MATERIAL AND METHODS: This retrospective study enrolled 53 NMLs (31 benign, 22 malignant). Two radiologists (readers 1 and 2 had 15 years and 1 year of experience in breast imaging, respectively) independently reviewed each study and determined the BI-RADS category using US alone and again after adding SWE to US. Diagnostic performances of US alone and US combined with SWE were compared for both readers. Areas under receiver operating characteristic curves (AUCs) were estimated. The levels of interobserver agreement were determined by the calculated kappa coefficient. RESULTS: With the addition of SWE to US, AUCs for differentiation of benign and malignant lesions increased significantly for the less experienced reader (0.56 vs. 0.79; p=0.028), but not for the more experienced reader (0.60 vs. 0.75; p=0.170). While evaluating US alone, the interobserver agreement was moderate, and the interobserver variability among the readers was statistically significant (k=0.493, p<0.001). After adding SWE, the agreement increased to 0.773, and the interobserver variability among the readers became non-significant (k=0.773, p=0.688) CONCLUSIONS: SWE increased the diagnostic performance of relatively less experienced reader significantly. SWE improved interobserver agreement of two readers with different levels of experience and reduced the interobserver variability in differentiating benign and malignant NMLs of the breast. KEY WORDS: Breast, Elastography, Ultrasound.
AIM: The aim of this study is to retrospectively assess the additional diagnostic value of shear wave elastography (SWE) added to ultrasound (US) versus US alone in differentiating malignant and benign non-mass lesions (NMLs) of the breast by readers with different experience levels and to assess interreader agreement. MATERIAL AND METHODS: This retrospective study enrolled 53 NMLs (31 benign, 22 malignant). Two radiologists (readers 1 and 2 had 15 years and 1 year of experience in breast imaging, respectively) independently reviewed each study and determined the BI-RADS category using US alone and again after adding SWE to US. Diagnostic performances of US alone and US combined with SWE were compared for both readers. Areas under receiver operating characteristic curves (AUCs) were estimated. The levels of interobserver agreement were determined by the calculated kappa coefficient. RESULTS: With the addition of SWE to US, AUCs for differentiation of benign and malignant lesions increased significantly for the less experienced reader (0.56 vs. 0.79; p=0.028), but not for the more experienced reader (0.60 vs. 0.75; p=0.170). While evaluating US alone, the interobserver agreement was moderate, and the interobserver variability among the readers was statistically significant (k=0.493, p<0.001). After adding SWE, the agreement increased to 0.773, and the interobserver variability among the readers became non-significant (k=0.773, p=0.688) CONCLUSIONS: SWE increased the diagnostic performance of relatively less experienced reader significantly. SWE improved interobserver agreement of two readers with different levels of experience and reduced the interobserver variability in differentiating benign and malignant NMLs of the breast. KEY WORDS: Breast, Elastography, Ultrasound.