High diagnostic accuracy of non-restricted diffusion in breast DWI to predict benign BI-RADS category 4 lesions can avoid unnecessary breast biopsies

Introduction: Breast diffusion-weighted imaging (DWI) can reclassify a significant number of suspicious breast images as benign, thus avoiding unnecessary biopsies. This study evaluated the diagnostic performance of breast DWI in differentiating benign and malignant BI-RADS category 4 lesions with histopathologic confirmation. Material and Methods: This crosssectional study included women with a BI-RADS category 4 breast lesion with suspected malignancy on mammography and/or ultrasound. BI-RADS subcategories 4a, 4b, and 4c were recorded. Breast DWI was performed to assess nonrestricted and restricted diffusion. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy of breast DWI for predicting benign and malignant lesions were calculated. Percutaneous breast biopsy was the gold standard. Results: Seventy-three women with a mean age of 49.4 ± 13.9 years with BI-RADS category 4 lesions were included. There were 48 benign, 2 benign with upgrade potential (BWUP), and 23 malignant lesions with histopathologic confirmation. Most of the benign lesions (n = 41, 85.4%) showed non-restricted diffusion, while only 7 (14.6 %) showed restricted diffusion (p < 0.001). In contrast, all of the malignant lesions (n = 23, 100%) showed restricted diffusion (p < 0.001). Non-restricted diffusion had a sensitivity of 85.4% and a specificity of 96.0% for predicting benign BI-RADS category 4 lesions. The PPV was 97.6% and the NPV 77.4%. The diagnostic accuracy was 89.0%. Conclusion: Our study shows that non-restricted diffusion in breast DWI has a high diagnostic accuracy in predicting benign BI-RADS category 4 breast lesions. This improved lesion characterization by breast DWI can reduce the number of false positive results and unnecessary biopsies in benign breast lesions.