Tumour-related factors and prognosis in breast cancer detected by screening

Abstract

Background: Breast cancer detected by screening has an unexplained prognostic advantage beyond stage shift compared with cancers detected clinically. The aim was to investigate biological factors in invasive breast cancer, with reference to mode of detection and rate of death from breast cancer.

Methods: Histology, oestrogen receptor α and β, progesterone receptor, human epidermal growth factor receptor (HER) 2, cyclin D1, p27, Ki-67 and perinodal growth were analysed in 466 tumours from a prospective cohort, the Malmö Diet and Cancer Study. Using logistic regression, odds ratios were calculated to investigate the relationship between tumour characteristics and mode of detection. The same tumour factors were analysed in relation to standard prognostic features. Death from breast cancer was analysed using Cox regression with adjustments for standard tumour factors; differences following adjustment were analysed by means of Freedman statistics.

Results: None of the biological tumour characteristics varied with mode of detection of breast cancer. After adjustment for age, tumour size, axillary lymph node involvement (ALNI) and grade, women with cancer detected clinically had an increased risk of death from breast cancer (hazard ratio 2·48, 95 per cent confidence interval 1·34 to 4·59), corresponding to a 37·2 per cent difference compared with the unadjusted model. Additional adjustment for biological tumour factors studied caused only minor changes.

Conclusion: None of the biological tumour markers investigated explained the improved prognosis in breast cancer detected by screening. None of the factors was related to ALNI, suggesting that other mechanisms may be responsible for tumour spread.