Comparison of interval breast cancers with 2D digital mammography versus 3D digital breast tomosynthesis in a large community-based practice.

Incidence of interval cancers is an important outcome in assessing efficacy of screening. Our primary objective was to compare the incidence of interval cancers detected with two-dimensional digital mammography (DM) versus digital breast tomosynthesis (DBT) in a large community health system. Our secondary objectives were to compare the patient and tumor characteristics of interval cancers, cancer detection rate, and recall rate. Interval cancers before and after implementation of DBT (2012-2014 DM group; 2016-2018 DBT group) were reviewed. Patient factors (age, race, breast density, personal history of breast cancer, family history of breast cancer, known BRCA-1 or BRCA-2 genetic mutation, baseline mammogram, and presentation) and tumor characteristics (in situ versus invasive, grade, size, hormone receptor status, and nodal status) were compared with the chi-squared test or the MidP exact test. Rates (detection and recall) were compared using a z-score. The rates of interval cancers with DM (0.30 per 1000 [35/117 099]) and DBT (0.33 per 1000 [40/119 746]) were similar (P = .3). Proportion of node-positive interval cancers was lower in the DBT group (22.9% [8/35] vs 48% [15/31], p.01). Otherwise, the patient and tumor characteristics were similar. The cancer detection rate increased (5.9 per 1000 [709/119 746] vs 3.5 per 1000 [411/117 099], P = .0001), and the recall rate decreased with DBT (8.6% [10 347/119 746] versus 10.7% [12 508/117 099], (P < .0001). Although the cancer detection rate was higher with DBT, the rate of interval breast cancers was similar in both groups. Node-positive invasive interval cancers were decreased with DBT.