An individual participant data meta-analysis of breast cancer detection and recall rates for digital breast tomosynthesis versus digital mammography population screening

Background: Although digital breast tomosynthesis (DBT) improves breast cancer screen-detection compared to digital mammography (DM), there is less evidence on comparative screening outcomes by age and breast density, and inconsistent evidence on its effect on recall rate. Method: We performed an individual participant data (IPD) meta-analysis from DBT screening studies (identified to November, 30 2019) that contributed to the study protocol. We estimated and compared cancer detection rate (CDR), recall rate, and positive predictive value (PPV) for recall for DBT and DM screening. Two-stage random-effects meta-analyses of detection outcomes adjusted for study and age, and were estimated in age and density subgroups. Screen-detected cancer characteristics were summarized descriptively within studies and screening-groups. Results: Four prospective studies, from European population-based programs, contributed IPD for 66,451 DBT-screened participants and 170,764 DM-screened participants. Age-adjusted pooled CDR difference between DBT and DM was 25.49 of 10,000 (95% CI:6.73-44.25). There was suggestive evidence of a higher CDR for DBT compared to DM in the high-density (35.19 of 10,000; 95% CI:17.82-56.56) compared to low-density (17.4 of 10,000; 95% CI:7.62-27.18) group (P =.08). Pooled CDR difference between DBT and DM did not differ across age-groups (P =.71). Age-adjusted recall rate difference was 0.18% (95% CI:-0.80–1.17), indicating no difference between DBT and DM- this finding did not differ across age-groups (P =.96). Recall PPV was higher for DBT than DM with an estimated rate ratio of 1.31 (95% CI:1.07-1.61). Discussion: DBT improved CDR compared to DM in all age and density groups. DBT also had higher recall PPV than DM, although further research is needed to explore the heterogeneity in recall rates across studies.