Sonoelastography for 1786 non-palpable breast masses: diagnostic value in the decision to biopsy

Abstract

To evaluate the diagnostic value of sonoelastography by correlation with histopathology compared with conventional ultrasound on the decision to biopsy. Prospectively determined BI-RADS categories of conventional ultrasound and elasticity scores from strain sonoelastography of 1786 non-palpable breast masses (1,523 benign and 263 malignant) in 1,538 women were correlated with histopathology. The sensitivity and specificity of two imaging techniques were compared regarding the decision to biopsy. We also investigated whether there was a subset of benign masses that were recommended for biopsy by B-mode ultrasound but that had a less than 2% malignancy rate with the addition of sonoelastography. The mean elasticity score of malignant lesions was higher than that of benign lesions (2.94 ± 1.10 vs. 1.78 ± 0.81) (P < 0.001). In the decision to biopsy, B-mode ultrasound had higher sensitivity than sonoelastography (98.5% vs. 93.2%) (P < 0.001), whereas sonoelastography had higher specificity than B-mode ultrasound (42.6% vs. 16.3%) (P < 0.001). BI-RADS category 4a lesions with an elasticity score of 1 had a malignancy rate of 0.8%. Sonoelastography has higher specificity than B-mode ultrasound in the differentiation between benign and malignant masses and has the potential to reduce biopsies with benign results. • Sonoelastography has higher specificity than B-mode ultrasound in distinguishing benign from malignant masses. • Sonoelastography could potentially help reduce the number of biopsies with benign results. • Lesion stiffness on sonoelastography correlated with the malignant potential of the lesion.