Using tumor phenotype, histological tumor distribution, and mammographic appearance to explain the survival differences between screen-detected and clinically detected breast cancers.

Abstract

In the era of mass screening for breast cancer with mammography, it has been noted that conventional tumor attributes and mammographic appearance are insufficient to account for the better prognosis of screen-detected tumors. Such prognostication may require additional updated pathological information regarding tumor phenotype (e.g., basal status) and histological tumor distribution (focality). We investigated this hypothesis using a Bayesian approach to analyze breast cancer data from Dalarna County, Sweden. We used data for tumors diagnosed in the Swedish Two-County Trial and early service screening period, 1977-1995, and from the mature service screening period, 1996-1998. In the early period of mammographic screening (1977-1995), the crude hazard ratio (HR) of breast cancer death for screen-detected cases compared with symptomatic ones was 0.22 (95% CI: 0.17-0.29) compared with 0.53 (95% CI: 0.34-0.76) when adjusted for conventional tumor attributes only. Using the data from the mature service screening period, 1996-1998, the HR was 0.23 (95% CI: 0.08-0.44) unadjusted and 0.71 (95% CI: 0.26-1.47) after adjustment for tumor phenotype, mammographic appearance, histological tumor distribution, and conventional tumor attributes. The area under the ROC curve (AUC) for the prediction of breast cancer deaths using these variables without the detection mode was 0.82, only slightly less than that observed when additionally including the detection mode (AUC=0.83). Using Freedman statistics, conventional tumor attributes and mammographic appearances explained 58% (95% CI: 57.5-58.6%) of the difference of breast cancer survival between the screen-detected and the clinically detected breast cancers, whereas the corresponding figure was increased to 77% (95% CI: 75.6-77.6%) when adding the two information on tumor phenotype and histological tumor distribution. The results indicated that conventional tumor attributes and mammographic appearance are not sufficient to be interim markers for explaining the survival difference between screen-detected and clinically detected cancers in the era marked by the widespread use of mammography. Additional information on tumor phenotype and histological distribution may be added as effective interim markers for explaining the benefit of the early detection of breast cancer with mammography.