Evolution of sites of recurrence after early breast cancer over the last 20 years: implications for patient care and future research.

Improvements in adjuvant therapy have led to a sustained fall in recurrences after early breast cancer. The differential reduction in local and systemic recurrences is poorly understood. This study aimed to explore changes in the distribution of loco-regional and distant recurrences in clinical trials reported over the last 20 years. We also aimed to determine the relative impact of adjuvant chemotherapy and endocrine therapy. MEDLINE search for adjuvant, phase III randomized breast cancer clinical studies between January 1990 and March 2011 was performed. Neoadjuvant, single agent biologics and studies that did not report the proportion of loco-regional and distant recurrences were excluded. The change in the frequency of recurrences was assessed as the nonparametric correlation between the number of loco-regional recurrences as a proportion of all recurrences and time. Studies were weighted by sample size. The differential effect of chemotherapy and endocrine therapy was also assessed. Fifty-three randomized clinical trials with a total of 86,598 patients were included in the analysis. Between 1990 and 2011, the proportion of loco-regional recurrences has decreased from approximately 30 to 15 % (Spearman's ρ = -0.40, p < 0.001). There was no interaction between type of surgery (mastectomy vs. lumpectomy), administration of adjuvant radiation therapy and menopausal status and the correlation of loco-regional recurrences and time. Chemotherapy regimen showed a larger negative correlation compared with endocrine therapy ( ρ = 0.49 vs. ρ = 0.24). Advances in treatment of early breast cancer have differentially reduced the proportion of loco-regional recurrences compared with distant recurrences. In recent trials, loco-regional recurrences account for less than 10-15 % of all recurrences. These falling event rates may affect patient care, especially when deciding on treatments influencing loco-regional control. This change may also impact on the design of clinical trials assessing loco-regional therapy such as surgery and/or local radiation therapy.