Breast strain elastography: Observer variability in data acquisition and interpretation

Abstract

Objective: To analyze the observer reproducibility of breast strain elastography in elasticity image acquisition and elasticity image interpretation.

Methods: This was an institutional ethics committee approved prospective study. One hundred twenty-four breast lesions in 118 women (mean age 45.39 ± 12.97 years old, range 21-77 years old) were examined with strain elastography by two blinded radiologists in turn. Three blinded observers separately reviewed and recorded the elasticity score of each lesion obtained by the two performers. The interobserver reproducibility of elasticity image acquisition between the two performers, the interobserver and intraobserver reproducibility of elasticity image interpretation among observers were evaluated. The diagnostic performance of strain elastography was compared between the two performers.

Results: Fifty-three lesions were malignant and 71 were benign. The interobserver kappa value was 0.438 for the elasticity score between the two performers. Between the three observers, the overall interobserver and intraobserver kappa value was 0.365 and 0.655, respectively. There was no significant difference of the area under the receiver operator characteristic curve (Az) value for the elasticity score between performer 1 and 2 (P = 0.143).

Conclusions: Our results suggested moderate interobserver reproducibility in breast strain elasticity image acquisition, poor interobserver and good intraobserver agreement in image interpretation.

Keywords: Breast lesion; Elasticity score; Elastography; Ultrasound.