Comparison of BI-RADS categories and recommendations from mammography alone and mammography with subsequent breast ultrasound in women with BI-RADS 1 and 2 screening mammogram

Abstract

Objective : To evaluate false negative rate of mammography according to BI-RADS categories and recommendations before breast ultra-sound (US) in women with BI-RADS 1 and 2 screening mammo-graphy and to study the effects of breast composition, age and availability of prior mammography on changing of recommendation after breast US.

Setting : Bangkok Metropolitan Administration General Hospital

Research design : A retrospective cohort study.

Patients : Women who received screening mammography with mammographic BI-RADS 1 and 2, and subsequent breast US were performed in the same day from July 2005 to February 2007.

Methods : Mammograms of 1003 women were reviewed. Mammographic findings were recorded and mammographic BI-RADS category (BI-RADSJ and recommendation (RECOMJ were assigned. Breast US reports and images were reviewed. US findings, final BI-RADS category (BI-RADS1) and final recommendation (RECOM were recorded. False negative rate of BI-RADS 1 and RECOM_ were analyzed using BI-RADS and RECOM as the gold standards. The effects of age, breast composition and availability of previous mammogram on incorrect recommendation were evaluated.

Results : False negative rate of BI-RADSM categories was 7.68% (95% Cl = 6.14- 9.56) and ofRECOMM was 2.59% (95%CI = 1.73- 3.83): The effects of breast composition and availability of prior mammogram for comparison on recommendation change after breast US were statistically significant. Women with dense breasts were 10.64 times (95%CI = 1.43 - 79.13) more likely to have recommendation change than women with non-dense breasts. Women without prior mammogram were 11.44 times (95%CI = 2.68 - 48.78) more likely to have recommendation change than women with prior films. Age had no significant effect on recommendation change • after breast US.

Conclusion : A small proportion (2.59%) of women with mammographic BI-RADS 1 and 2 in screening mammography had recommendation change following breast US, particularly women with non-dense breasts and women with prior mammography available for comparison. This suggests lack of necessity to perform breast US in women with negative or benign findings on screening mammography of non-dense breast group and also in the group with available interval mammographic images.