Cancer Causes & Control

, Volume 18, Issue 10, pp 1187–1198

Qualitative age interactions (or effect modification) suggest different cancer pathways for early-onset and late-onset breast cancers

  • William F. Anderson
  • Bingshu E. Chen
  • Louise A. Brinton
  • Susan S. Devesa
Original Paper

DOI: 10.1007/s10552-007-9057-x

Cite this article as:
Anderson, W.F., Chen, B.E., Brinton, L.A. et al. Cancer Causes Control (2007) 18: 1187. doi:10.1007/s10552-007-9057-x

Abstract

Background

Prior to 1999–2000, breast cancer incidence rates had risen for decades, though more among older than younger women.

Materials and methods

To further explore the impact of advancing age-at-diagnosis upon breast cancer incidence, we used the National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) program (1974–2003).

Results

Over time, we observed age interactions by tumor grade, stage, and race. For example, among women ages <40 years, high-grade lesions were more common than low-grade tumors for all time periods. Among women ages 40+ years, high-grade lesions were more common during early time periods then trend lines crossed, after which low-grade tumors were more common than high-grade lesions. Notably, the transition (crossover point) occurred earlier with advancing age-at-diagnosis.

Conclusion

The reversal (crossing) of incidence rates from high to low-grade tumors among women 40+ years is a qualitative age interaction, probably due to changing age-related risk factor and/or screening patterns, where mammography preferentially detected tumors of low malignant potential among older women. Though once thought to be rare or artifactual, qualitative age interactions suggest breast cancer heterogeneity. Indeed, if real, qualitative age interactions (effect modifications) imply different etiologic pathways for early-onset and late-onset types of breast cancer.

Keywords

Breast cancer incidence Risk factors Breast cancer etiology SEER Age-at-diagnosis Temporal trends 

Copyright information

© Springer Science + Business Media B.V. 2007

Authors and Affiliations

  • William F. Anderson
    • 1
  • Bingshu E. Chen
    • 1
  • Louise A. Brinton
    • 2
  • Susan S. Devesa
    • 1
  1. 1.Biostatistics BranchDHHS/NIH/NCI/DCEGBethesdaUSA
  2. 2.Hormonal and Reproductive Epidemiology BranchDHHS/NIH/NCI/DCEGBethesdaUSA

Personalised recommendations