Background: Cancer Detection Rate (CDR) and Interval Cancer Rate (ICR) are key parameters for evaluating breast cancer screening performance. Interval cancers (IC) are generally more advanced and aggressive than screen-detected breast cancers (SDBC). It is often assumed that higher CDR should be associated with lower ICR, and that a combination of high CDR and low ICR could reflect greater ability to prevent breast cancer death. However, the relationship between CDR and ICR, particularly for advanced and aggressive interval cancers, remains unclear.
Objectives: This systematic review with meta-analysis assessed the association between CDR and ICR across breast screening studies, considering tumour characteristics and screening technology.
Methods: We selected cohort studies conducted in general female populations aged 40-75 years. CDR was defined as the number of SDBC per 1,000 screenings. ICR was modelled as an incidence rate per 1,000 person-time to account for heterogeneous intervals (1-3 years) between screening rounds. Descriptive statistics were summarized through medians and interquartile ranges (IQR). Associations were examined using negative binomial generalized linear mixed models (GLMMs), with a random intercept to account for within-study clustering; results were expressed as incidence rate ratios (IRR with 95% CI). Analyses were conducted overall and stratified by screening technology (screen-film mammography (SFM), digital mammography (DM), and digital breast tomosynthesis (DBT)). When available, analyses stratified by tumour grade (G1-G3) and lymph node involvement were performed. Sensitivity analyses excluded outliers, restricted analyses to studies with two-year follow-up, and incorporated select studies initially excluded.
Results: 78 studies conducted between 1992 and 2025 in different countries were selected, providing 111 observations and a total of 284,882 SDBC and 110,667 IC. The summary estimate of CDR was 5.95 (IQR: 5.47, 6.46). CDR differed significantly by screening technologies (p<0.001), with DBT showing the highest CDR, and across countries (p=0.002), with European studies exhibiting higher CDRs than those from Asia and North America. The summary estimate of ICR was 0.76 (IQR: 0.69, 0.83). ICR did not differ significantly by technology (p=0.34) or screening interval (p=0.27). Overall, no consistent association between CDR and ICR was identified: IRR = 1.01 (95%CI: 0.98, 1.04; p=0.49). When considering each technology separately, increasing CDR was associated with slight increases in ICR across all modalities, with IRRs of 1.05 (95%CI: 0.99, 1.11), 1.19 (95%CI: 1.08, 1.31), and 1.06 (95%CI: 0.96, 1.16) for SFM, DM, and DBT, respectively (Figure 1). No clear associations were found also stratifying by tumour grade or lymph node involvement. Increasing CDR was not associated with lower ICR of in situ (p=0.88) or lymph node positive (p=0.33) cancers; increasing CDR of grade 1-2 cancers was not associated with lower ICR of grade 3 cancers (p=0.22). Findings were robust across sensitivity analyses.
Conclusions: This meta-analysis found no consistent relationship between CDR and ICR in breast cancer screening, calling into question the assumption that greater cancer detection can lead to lower occurrence of IC. Specifically, higher CDR was not associated with lower overall ICR or with lower ICR of in situ, grade 3, or lymph node-positive cancers.

Association between CDR and ICR, overall and by screening technology