Interval breast cancer rates for tomosynthesis vs mammography population screening: a systematic review and meta-analysis of prospective studies

Objectives

We aimed to synthesise evidence from prospective studies of digital breast tomosynthesis (DBT) screening to assess its effectiveness compared to digital mammography (DM). Specifically, we examined whether DBT reduces interval cancer rates (ICRs) in population breast cancer screening.

Materials and methods

We performed a systematic review and meta-analysis of DBT screening studies (identified from January 2013 to March 2024). We included both RCTs and non-randomised prospective studies that used an independent comparison for our primary outcome ICRs. The risk of bias was assessed with QUADAS-2. We compared the ICR, cancer detection rate (CDR), and recall rate of DBT and DM screening using random effects meta-analysis models. Subgroup analyses estimated outcomes by study design. Sensitivity analyses estimated absolute effects from relative effects.

Results

Ten prospective studies (three RCTs, seven non-randomised) were eligible; all had a low risk of bias. There were 205,245 DBT-screened and 306,476 DM-screened participants with follow-up for interval cancer data. The pooled absolute ICR did not significantly differ between DBT and DM: −2.92 per 10,000 screens (95% CI: −6.39 to 0.54); however subsequent subgroup analysis indicated certain study designs may have biased this ICR estimate. Pooled ICR from studies that only sampled groups from the same time and region indicated DBT led to 5.50 less IC per 10,000 screens (95% CI: −9.47 to −1.54). Estimates from subgroup analysis that compared randomised and non-randomised trials did not significantly differ.

Conclusion

This meta-analysis provides suggestive evidence that DBT decreases ICR relative to DM screening; further evidence is needed to reduce uncertainty regarding ICR differences between DBT and DM.

Key Points

Question Does DBT have long-term benefits over standard DM?

Finding We find suggestive evidence in our primary analysis and stronger evidence in a follow-up analysis that DBT reduces interval cancers.

Clinical relevance This meta-analysis provides the first indication that DBT may detect additional cancers that are clinically meaningful, based on suggestive evidence of a reduction in ICR. This finding does not preclude the simultaneous possibility of overdiagnosis.

Graphical Abstract