Differentiation of benign from malignant nonpalpable breast masses: a comparison of computer-assisted quantification and visual assessment of lesion stiffness with the use of sonographic elastography.

BACKGROUND: Elastography has shown potential in differentiating benign from malignant breast tumors, but interobserver variability between experienced and inexperienced readers limits its wide usage.
PURPOSE: To compare the diagnostic performance of computer-assisted quantification and visual assessment of lesion stiffness with the use of sonographic elastography for the differentiation of benign from malignant nonpalpable breast masses.
MATERIAL AND METHODS: Sonographic elasticity images of 120 nonpalpable breast masses (70 benign and 50 malignant masses) were obtained in 120 women prior to performing a core biopsy. After subtraction of B-mode images from color elasticity images, the mean strain value of the lesion was computed. Elasticity images were also reviewed and were assigned a score on a five-point scale by two breast radiologists in consensus. Results were evaluated by using receiver operating characteristic (ROC) curve analysis.
RESULTS: The mean +/- standard deviation values of strain were 221+/-18 for malignant lesions and 175+/-21 for benign lesions (P<0.001). For the elasticity score, the mean score was 3.5+/-0.1 for the malignant masses and 2.0+/-0.9 for the benign masses (P<0.001). The overall Pearson's correlation coefficient between the strain values and elasticity score was 0.689 (P<0.001). The area under the ROC curve (A(z)) value was 0.878 for use of the computer-assisted quantification method and 0.850 for visual assessment by the radiologists. The difference was not statistically significant (P=0.198).
CONCLUSION: Computer-assisted quantification and visual assessment of lesion stiffness with the use of sonographic elasticity images had comparable diagnostic performance for the differentiation of nonpalpable breast masses.