Diagnostic Performance of Digital Breast Tomosynthesis, Unenhanced MRI, and Their Combination in the Preoperative Assessment of Breast Cancer: A Multi-reader Study

Rationale and Objectives: To compare the diagnostic performance of digital breast tomosynthesis (DBT) and unenhanced magnetic resonance imaging (UMRI) in the preoperative assessment of breast cancer. Materials and Methods: We retrospectively included 59 patients with 74 pathology-proven cancers who underwent DBT and preoperative 1.5 T magnetic resonance imaging between January 2016 and February 2017. Four residents with 2-3 years of experience, blinded to pathology, independently reviewed DBT and UMRI (diffusion-weighted and unenhanced T1-weighted sequences), using the breast imaging reporting and data system (BI-RADS) and a 0-5 Likert score, respectively. We calculated per-lesion sensitivity and positive pre-dictive value of DBT, UMRI, and combined DBT+UMRI, as well as the agreement between DBT and UMRI vs. pathology in assessing cancer size (Bland-Altman analysis). Logistic regression was performed to assess clinical features predictive of missing cancer. Results: Of 74 lesions, 84% were invasive ductal carcinoma, 27% of which with an in situ component; 31% of cancers were <10 mm large. Sensitivity of UMRI (74-85%) was equal or higher than that of DBT (68-82%), with similar positive predictive value (93-97% vs. 98-100%, respectively). DBT+UMRI increased the sensitivity up to 88%. UMRI showed closer limits of agreement with pathological size than DBT. Missing cancer was independently predicted by size <10 mm on DBT, UMRI, and DBT+UMRI (odds ratio 18.7, 5.1, and 13.3, respectively), and by increased breast density on DBT alone (odds ratio 3.50). Conclusion: UMRI was equal or better than DBT in the preoperative assessment of breast cancer. Combined imaging achieved up to 88% per-lesion sensitivity, suggesting potential use in clinical practice. (C) 2020 The Association of University Radiologists. Published by Elsevier Inc. All rights reserved.