Romuald Ferre, Thad Benefield, Cherie M Kuzmiak
Across more than 1.4 million screening examinations, DBT was not associated with a meaningful reduction in invasive interval breast cancer incidence compared with DM. These findings suggest that improved mammographic lesion conspicuity with DBT may not substantially affect cancers that are biologically aggressive, rapidly developing, or mammographically occult.
RATIONALE AND OBJECTIVES: To evaluate whether digital breast tomosynthesis (DBT) reduces the incidence of invasive interval breast cancers compared with digital mammography (DM) in population-based breast screening studies, and whether this effect differs by screening interval.
MATERIALS AND METHODS: We performed a systematic review and meta-analysis of comparative screening studies evaluating invasive interval breast cancers after negative screening examinations with DBT versus DM. Eligible studies reported modality-specific denominators and invasive interval cancers diagnosed after a negative screening examination and before the next scheduled screening round. The primary effect measure was the absolute risk difference, defined as DBT minus DM. Pooled estimates were calculated using Mantel-Haenszel common-effect and inverse-variance random-effects models with Paule-Mandel estimation of between-study variance. Prespecified subgroup analyses were performed by study design and intended screening interval, classified as annual or nonannual.
RESULTS: Thirteen studies were included, comprising 1407,624 screening examinations, including 548,368 DBT examinations and 859,256 DM examinations, with 4370 invasive interval breast cancers. Three studies were classified as annual screening studies and 10 as nonannual screening studies. Overall, DBT was not associated with a statistically significant reduction in invasive interval breast cancer incidence compared with DM. Results were consistent across randomized trials and observational cohorts. Descriptive subgroup analyses by screening interval did not suggest a meaningful difference in interval cancer incidence between DBT and DM.
CONCLUSION: Across more than 1.4 million screening examinations, DBT was not associated with a meaningful reduction in invasive interval breast cancer incidence compared with DM. These findings suggest that improved mammographic lesion conspicuity with DBT may not substantially affect cancers that are biologically aggressive, rapidly developing, or mammographically occult.