Article Sidebar
Main Article Content
Abstract
Background: Breast cancer remains a major health concern among women in Libya, with many cases presenting at a relatively young age. The Breast Imaging Reporting and Data System (BI-RADS) provides a standardized approach for classifying breast lesions; however, its diagnostic performance in the Libyan population has not been widely evaluated. Objective: To evaluate the performance of the BI-RADS classification system in grayscale ultrasonography and mammography for differentiating benign and malignant breast lesions among Libyan women, in correlation with histopathological results Methods: This retrospective study included 274 women who underwent breast imaging followed by histopathological confirmation between 2016 and 2017 at NCI Sabratha. BI-RADS categories were compared with final histopathology results. Diagnostic performance was calculated using sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), accuracy, and receiver operating characteristic (ROC) analysis. Results: BI-RADS demonstrated high sensitivity (88.2%), strong NPV (91.6%), and good overall accuracy (80.7%). However, several categories showed deviations from expected malignancy rates. BI-RADS 2 and 3 exhibited higher-than-expected malignancy rates (2.8% and 10.3%, respectively). BI-RADS 4A and 4B demonstrated markedly elevated malignancy rates (44.2% and 71.4%), while BI-RADS 4C aligned with international benchmarks (95.6%). BI-RADS 5 showed a lower-than-expected malignancy rate (78.4%). Conclusion: The BI-RADS system provides strong overall diagnostic performance in this cohort of Libyan women; however, discrepancies in malignancy rates across several categories highlight variability in interpretive accuracy. Enhanced radiologist training and quality assurance programs are recommended.