Differential diagnosis of mammographically and clinically occult breast lesions on diffusion‐weighted MRI

Abstract

PURPOSE: To investigate the diagnostic performance of diffusion-weighted imaging (DWI) for mammographically and clinically occult breast lesions. MATERIALS AND METHODS: The study included 91 women with 118 breast lesions (91 benign, 12 ductal carcinoma in situ [DCIS], 15 invasive carcinoma) initially detected on dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) and assigned BI-RADS category 3, 4, or 5. DWI was acquired with b = 0 and 600 s/mm(2). Lesion visibility was assessed on DWI. Apparent diffusion coefficient (ADC) values were compared between malignancies, benign lesions, and normal (no abnormal enhancement on DCE-MRI) breast tissue, and the diagnostic performance of DWI was assessed based on ADC thresholding. RESULTS: Twenty-four of 27 (89%) malignant and 74/91 (81%) benign lesions were hyperintense on the b = 600 s/mm(2) diffusion-weighted images. Both DCIS (1.33 +/- 0.19 x 10(-3) mm(2)/s) and invasive carcinomas (1.30 +/- 0.27 x 10(-3)mm(2)/s) were lower in ADC than benign lesions (1.71 +/- 0.43 x 10(-3)mm(2)/s; P < 0.001), and each lesion type was lower in ADC than normal tissue (1.90 +/- 0.38 x 10(-3)mm(2)/s, P <or= 0.001). Receiver operating curve (ROC) analysis showed an area under the curve (AUC) of 0.77, and sensitivity = 96%, specificity = 55%, positive predictive value (PPV) = 39%, and negative predictive value (NPV) = 98% for an ADC threshold of 1.60 x 10(-3)mm(2)/s. CONCLUSION: Many mammographically and clinically occult breast carcinomas were visibly hyperintense on diffusion-weighted images, and ADC enabled differentiation from benign lesions. Further studies evaluating DWI while blinded to DCE-MRI are necessary to assess the potential of DWI as a noncontrast breast screening technique.