2120. Rates and Characteristics of False-Negative Cancers With Digital Breast Tomosynthesis
Authors * Denotes Presenting Author
  1. Manisha Bahl *; Harvard Medical School; Massachusetts General Hospital
The impact of digital breast tomosynthesis (DBT) on long-term patient outcomes is not known, but the false-negative rate in women undergoing DBT is considered to be a surrogate marker for long-term outcomes. The purpose of this study is to determine the false-negative rate of breast cancer screening with DBT and to evaluate the characteristics of interval cancers.

Materials and Methods:
In this Institutional Review Board-approved and Health Insurance Portability and Accountability Act-compliant study, consecutive screening DBT examinations from January 2016 to December 2018 were retrospectively reviewed. Cancers were considered false-negative cancers if diagnosed within 365 days of a negative screening DBT examination. Medical records were reviewed for mode of presentation, biopsy pathology results, and surgical pathology results. Characteristics of symptomatic false-negative cancers and MRI-detected false-negative cancers were compared using the Pearson’s chi-squared test.

The false-negative rate was 0.8 per 1000 screening examinations (110/143,622). Of the 110 women (mean age 59 years, range 41-90 years) with false-negative cancers, 32.7% (36/110) had a prior history of breast cancer, and 67.3% (74/110) had heterogeneously dense or extremely dense breasts on mammography. More than 85% (n = 96) of the 110 false-negative cancers were invasive malignancies, which had a mean size of 15 mm (range 0.5-85 mm). Of those cases with available data, 81.3% (74/91) were grade 2 or 3, 84.0% (79/94) were estrogen receptor-positive, and 35.3% (30/85) were lymph node-positive. The false-negative cancers were diagnosed based on symptoms in 54.5% (60/110) of cases, detected on high-risk screening MRI examinations in 40.0% (44/110), incidentally detected on other imaging modalities (such as PET/CT) in 4.5% (5/110), and incidentally diagnosed at elective prophylactic mastectomy in 0.9% (1/110). Compared to symptomatic false-negative cancers, MRI-detected false-negative cancers were less likely to be invasive malignancies (75.0% [33/44] versus 96.7% [58/60], p < 0.001) and less likely to be lymph-node positive (19.4% [6/31] versus 42.0% [21/50], p = 0.04).

In our study cohort of 143,622 women who underwent breast cancer screening with DBT, the false-negative rate was 0.8 per 1000 screening examinations. MRI-detected cancers, which account for two-fifths of false-negative cancers, had favorable tumor characteristics compared to symptomatic false-negative cancers. Further research is needed to understand the role of supplemental screening with MRI in women at risk for interval cancers with DBT alone and to understand the impact of screening with DBT on long-term patient outcomes.