Send to

Choose Destination
Eur J Radiol. 2013 Aug;82(8):1231-5. doi: 10.1016/j.ejrad.2013.02.032. Epub 2013 Mar 27.

Simple rules for ultrasonographic subcategorization of BI-RADS®-US 4 breast masses.

Author information

Department of Obstetrics and Gynecology, Faculty of Medical Sciences, State University of Campinas - Unicamp, Campinas, São Paulo, Brazil.



To evaluate an objective method for ultrasonographic (US) subcategorization of BI-RADS(®)-US 4 breast masses based on clear and simple rules in order for woman to benefit from a more complete and homogeneous breast mass analysis.


In this cross-sectional study, we selected 330 women, with 339 US breast masses, classified as BI-RADS(®)-US 4. Three physicians experienced in breast imaging independently reviewed all US images, assessing mass shape, margins, orientation, echo texture and vascularity. These experts further subdivided the masses into subcategories 4a, 4b and 4c, according to simple US rules. Inter-observer agreement was calculated for US features categories and for final subcategory assessment. We also estimated the positive predictive value (PPV) for BI-RADS(®)-US subcategories 4a, 4b and 4c assigned by each of the three observers.


Pathological examination of all masses confirmed 144 (42%) malignant and 195 (58%) benign tumors. Moderate agreement was obtained for mass shape, margins, vascularity and for final BI-RADS(®)-US 4 subcategory. Substantial agreement was obtained for the description of mass orientation and echo texture. The PPV for subcategories 4a, 4b and 4c were, 17%, 45% and 85%, respectively, for the first observer and 20%, 38% and 79% and 17%, 40% and 85% for the other two observers.


Standardization of a US subcategorization of BI-RADS(®)-US 4 breast masses seems to be feasible, with substantial inter-observer agreement and progressive increase in the PPV in the subcategories 4a, 4b and 4c, provided that clear and simple classification rules are defined.

[Indexed for MEDLINE]

Supplemental Content

Full text links

Icon for Elsevier Science
Loading ...
Support Center