In a retrospective study, from August 1999 until June 2002, 6,999 consecutive mammograms were analyzed using Breast Imaging Reporting and Data System (BI-RADS™) guidelines, with the purpose of a regional analysis of mammographic findings in the categories 4 and 5 BI-RADS, which were all recommended to histopathologic study, evaluating the sensibility of this method in these categories, as an indicator of malignancy. The analysis revealed 43.65% (n = 3,055) of negative findings, 47.36% (n = 3,315) of benign lesions, 7.47% (n = 523) of probably benign lesions, 0.87% (n = 61) of suspicious lesions, and 0.64% (n = 45) of highly suspicious lesions. All the patients fitted in the categories 4 (61 patients) and 5 (45 patients) were referred for biopsy. The histopathologic analysis showed 27 patients in category 4 (44.26%) and 27 patients in category 5 (60%): 55.55% (n = 15) of category 4 patients showed malignant lesions; 96.29% (n = 26) of category 5 patients showed malignant lesions. The study demonstrates the high positive predictive value for malignancy in BI-RADS category 5 and the real need for follow-up investigation in BI-RADS category 4.

Keywords: BI-RADS, Mammography

We used the American College of Radiology BI-RADS™ (Breast Imaging Reporting and Data System) guidelines to perform a regional retrospective analysis of 4,968 mammograms. The frequency of each mammography finding was determined in order to categorize the mammograms into different groups. The analysis of the 4,968 mammograms revealed 44.4% (n = 2,208) of negative findings, 46.3% (n = 2,305) of benign lesions, 7.5% (n = 374) of probably benign lesions, 0.98% (n = 49) of suspicious lesions and 0.72% (n = 36) of highly suspicious lesions. This study also demonstrates the importance of using BI-RADS™ as a standardized report system for mammogram analysis, particularly as a predictor of malignancy and guideline for patient management.

Keywords: BI-RADS, Mammography