PURPOSE: The purpose of the study was to assess the final clinical outcome of BI-RADS Categories for diagnostic mammography and sonography. MATERIAL AND METHODS: We analysed 632 mammography and sonography examinations from women with diagnostic indications (age: 23 - 100, mean 58) performed during 2001 and 2003. All patients received mammography and sonography examinations at different outside facilities and all patients received an additional sonography examination at the university radiology department and if necessary supplemental mammographic views. Final clinical outcome (Histology: 554; follow-up: 78) was ascertained in each case and concordance of BI-RADS-categories for mammography and sonography and final diagnosis were analysed. RESULTS: Final diagnosis yielded 230 benign lesions (36 %) and 402 cancers (64 %). Concordance of BI-RADS Assessment and final outcome was documented in 542 cases (86 %). There were 11 correct category 1 and 2 assessments (2 %). 142 lesions were classified with BI-RADS 3 (22 %) with 5 false negative ratings. There were 264 category 4 lesions (42 %) with a PPV for a malignant lesion of 71 % (187/264) and finally 215 BI-RADS 5 lesions with a PPV of 98 % (210/215). Overall sensitivity of mammography was 92 % with specificity of 75 % and for sonography 86 % and 76 %. Mammography had a significantly higher detection rate for malignant lesions than sonography. The highest correlation between BI-RADS category and final outcome was documented for the diagnostic combination of mammography and sonography with a kappa-value of 0.817 (p <0.001), followed by mammography (kappa: 0.684) and sonography (kappa: 0.631). The overall correlation was 0.681 (p <0.001). CONCLUSION: BI-RADS assessments of diagnostic mammography and sonography yields in a high cancer detection rate with a justifiable part of false positive ratings.
