A Retrospective Analysis Of Screen Detected Breast Cancers And Interval Breast Cancers In A Population Based Breast Cancer Screening Programme (Breast Test Wales)

Abstract

Breast cancer is the most common malignant neoplasm and cause of death of women in developed countries. Breast cancers screening has been shown to reduce breast cancer mortality. Despite screening, breast cancers can occur between screening episodes, these are termed interval breast cancers (IC). These can be further divided into true interval cancers (Tl) - cancers that developed de novo since screening, false negative cancers (FN) - cancers missed on screening, occult cancers - cancers not visible on mammograms, and unclassified cancers - cancers where classifications were not possible. 

This study compared IC to screen detected cancers (SD) to assess histological differences, presentation of IC and its interval since screening, differences between Tl and FN, and evaluated the long term survival of IC and SD. 

The Welsh national database was accessed to identify all SD between 1998-2001 and all IC presented between 1998-2004 but attended screening between 1998-2001. Patient demographics, cancer histopathological reports, interval between screening and diagnosis of IC and all cause death data were extracted and analysed. 

The number of IC decreased with age (p<0.001). There were more tubular cancers diagnosed by screening than presented as interval cancers (p<0.001). IC was significantly larger in size compared to SD, but this was due to selection bias of the screening programme. IC had more nodal disease (p<0.001), higher cancer grades (p<0.001), and worse prognostic score (p<0.001). Ten-year survival was 82% for SD and 72% for IC (p<0.001). 

FN presented earlier than Tl (p=0.014), proportionally have less nodal disease (p<0.001), lesser cancer grades (p<0.001), and better prognostic score (p<0.001). Paradoxically, false negative cancers had worse long term survival (p<0.001). 

Reducing screening interval from 3 to 2 yearly would reduce IC rate by 36% but conferring no survival benefit. Survival of all cancers had improved significantly; NPI survival estimation requires revision to reflect latest data.