Ultrasonographic assessment of breast density

Abstract

Ultrasonographic (US) assessment of breast density has the potential to provide a nonionizing method. This study was to prospectively evaluate intermodality and interobserver agreements for assessment of breast density between US and mammography. Institutional review board approval was obtained. Forty-one women (mean 52.1 years; range 25-72 years) with variable breast density consented to participate. Eight radiologists blinded to mammographic information performed breast US for all participants and assessed each breast density using four categories based on the proportion of the breast occupied by the fibroglandular tissue. All participants underwent full-field digital mammography and mammographic density was independently assessed by eight radiologists 2 weeks after US using the breast imaging reporting and data system (BI-RADS) 4-category system. Intermodality agreements between US and mammographic assessments and interobserver agreements among radiologists were assessed using kappa statistics (D-0) and intraclass correlation coefficients (ICCs). There was substantial intermodality agreement between the US and mammographic assessments of breast density (D-0 = 0.65 and ICC = 0.80), and 68 % (222/328) of the assessments had exact agreement. When categories were dichotomized into fatty (categories 1 and 2) and dense (categories 3 and 4), 86 % (282/328) of the assessments had exact agreement (D-0 = 0.71). The interobserver agreement for the US assessments of breast density was substantial (average D-0 = 0.63, ICC = 0.82) and not significantly different from that for the mammographic assessments (average D-0 = 0.74, ICC = 0.85) (P = 0.701). US and mammography demonstrated substantial intermodality and interobserver agreement for assessment of breast density.