Can Breast Magnetic Resonance Imaging Prevent Biopsy or Change the Management of BI-RADS (R) Category 4 Breast Lesions?

Abstract

The BI-RADS (R) category 4 includes suspicious breast lesions which requires biopsy. The aim of this study is to investigate the contribution of breast magnetic resonance imaging to the management of BI-RADS (R) category 4 breast lesions detected by mammography and/or ultrasonography. Thirty-four lesions classified as BI-RADS (R) category 4A, 4B, or 4C by conventional methods were evaluated with magnetic resonance imaging. All lesions were coded by using the American College of Radiology BI-RADS (R) lexicon. Each lesion was verified with the result of pathology. Lesions were evaluated as BI-RADS (R) category 1 in 1 patient (2.9%), category 3 (20.6%) in 7 patients, category 4 in 25 patients (73.6%), and category 5 in 1 patient (2.9%) with breast magnetic resonance imaging. Only the BI-RADS (R) 4A lesion categories were changed by breast magnetic resonance imaging, and these lesions were pathologically diagnosed as benign. The negative predictive value of breast MRI for BI-RADS (R) category 4A lesions was calculated as 100%. In all BI-RADS (R) category 4 lesions, pathologically 4 lesions found to be high risk and 3 lesions found to be malignant. The sensitivity, specificity, positive predictive, and negative predictive value of breast magnetic resonance imaging in BI-RADS (R) category 4 lesions were calculated as 100, 29.6, 26.9, and 100%, respectively. The area under the ROC curve was calculated 0.648. Breast magnetic resonance imaging is promising to be used as a problem-solving modality in BI-RADS (R) category 4A breast lesions.