Objective: To determine what percentage of cancers, detected by screening ultrasonography (US), were detectable by full-field digital mammography (FFDM) and digital breast tomosynthesis (DBT).
Materials and methods: Eighty-four consecutive women in whom mammography was negatively interpreted and supplementary screening US initially detected breast cancers at outside hospitals underwent both FFDM and DBT. We excluded cases with overt suspicious findings on repeat mammography and ineligible cases. In the remaining 41 cases, three radiologists who were blinded to tumor location, even though they were aware that they had breast cancers independently reviewed both FFDM and DBT. The reference standard was the reference FFDM made by two unblinded reviewers who were aware of the tumor location and shape on DBT, US, and magnetic resonance imaging (MRI). The visibility score based on the correct marking was compared between FFDM and DBT.
Results: Among the 41 cases, the cancers were visible in 25 (61.0%) on FFDM and in 34 (82.9%) on DBT (P = 0.047) by the unblinded review. In the blinded analysis, the cancers were significantly more "constantly visible" in the three radiologists on DBT than on FFDM [53.7% (22/41) vs. 26.8% (11/41), respectively, P = 0.013]. The dominant lesion type was "focal asymmetry" on DBT (39.0%) and "asymmetry" on FFDM (31.7%).
Conclusions: Our analysis suggests that 54% of cancers that were detected by US and were not evident on 2D mammography were detectable by screening using DBT. Additional 29% of cancers were visualized on DBT, when the area of concern was known.
Keywords: Breast cancer; Comparison; Digital breast tomosynthesis; Full-field digital mammography; Screening; Ultrasonography.
Copyright © 2015 Elsevier Ltd. All rights reserved.