Breast symptoms and the risk of subsequent interval cancers and lethal breast cancers
Session type: Proffered paper
Theme: Epidemiology and prevention
Our aim was to estimate the association of breast symptoms reported at screen with subsequent interval breast cancers and lethal breast cancers.
The current study was nested within the follow up cohort of national breast cancer screening program using individual data available from screening registry (1992-2012) and cancer registry (1992-2014). Visits with (individual) breast symptoms (exposure group, n=213726) were frequency matched to non symptomatic visits to form a reference group (n=213726). The risk of interval cancers and lethal breast cancers were compared in those who reported symptoms with no symptomatic visits. Poisson regression model was used to estimate the risk at 95% confidence interval.
The risk of interval cancer was higher in those who reported lump (RR=3.48, 95% CI 3.16-3.82), retraction (RR=1.50, 95% CI 1.28-1.75) or secretion (RR= 2.65, 95% CI 2.04-3.37) compared to those with no reported symptom. In those interval cancers where lump was reported at index screen, the risk of dying from interval cancer was also higher (RR= 1.82, 95% CI 1.31-2.44) compared to those without lump. In addition, women having a lump also had a higher risk (RR =1.64, 95% CI 1.45-1.85) of cancer at next screening visit. Sensitivity of mammogram, as well as the overall screening episode sensitivity taking into account further assessment and diagnostic confirmation, was high in those who reported a symptom (for lump test sensitivity at 82.2% (95% CI 80.4-83.9) and episode sensitivity was at 78.5% (95% CI 76.7-80.3)).
The study showed higher risk of interval- and lethal breast cancers in women who reported symptoms at screen. Screening visits with symptoms, especially lump, requires detailed further assessment and/or re-reading of mammography findings to detect extra breast cancer cases and probably a shorter screening interval than visits without symptoms.