Impact of tumour size on axillary involvement and distant dissemination in breast cancer

BACKGROUND: Tumour size and nodal involvement are the two main prognostic factors in breast cancer (BC). Their impact on the natural history of BC is not fully captured by analyses that ignore their quantitative nature. METHOD: Data pertaining to 18 159 patients treated with primary surgery: 3661 at the Institut Gustave-Roussy (IGR, France) between 1954 and 1983, and 14 498 in the breast cancer registry in the Stockholm–Gotland Health Care region (SG, Sweden) between 1976 and 1999, were collected. The risks of distant metastases (DMs) and of nodal involvement were analysed according to tumour size with parametric models. RESULTS: Using SG 1976–1990 as the reference group, relative risks (RRs) for DM were equal to 1.42 (95% CI: 1.29–1.56; P<10(−10)) in IGR and 0.61 (95% CI: 0.55–0.67; P<10(−10)) in SG 1991–1999. Differences in tumour size explained the increased risk in IGR (RR adjusted for tumour size 1.09; 95% CI: 0.99–1.20; P=0.07), but not the decreased risk in SG 1991–1999 (adjusted RR: 0.63; 95% CI: 0.57–0.69; P<10(−10)). The relationship between tumour size and DM risk changed significantly during the 1990s. CONCLUSION: Early diagnosis is sufficient to explain differences in the prognosis before 1990. After 1990, the use of adjuvant systemic therapies is the main reason for the reduction in DM.