Introduction: The risk of breast cancer recurrence has been linked to tumour size, grade, oestrogen (ER) receptor status, and degree of lymph node (LN) involvement. However, the role of these variables in predicting time to relapse is not well defined. This study was designed to identify patient and primary tumour characteristics that predict risk periods for breast cancer recurrence within our institution, to enable more tailored surveillance strategies.
Methods: We retrospectively studied a cohort of 473 patients who presented to The Queen Elizabeth Hospital, Adelaide, Australia, with recurrent breast cancer between 1968 and 2008. Patient and primary tumour characteristics were collected, including age, menopausal status, tumour grade, size, ER and progesterone receptor (PR) status, and LN involvement and modeled against time to relapse using Kaplan-Meier survival curves.
Results: High tumour grade, size ≥ 20 mm, ER negativity, and PR negativity were shown on univariate analysis to correlate significantly with earlier recurrence (P < 0.0001, P = 0.0012, P = 0.0006, and P = 0.006). Multivariate analysis identified tumour grade and size as significant predictors of timing of relapse after adjustment for other variables. LN involvement, menopausal status, and age did not significantly correlate with earlier recurrence.
Conclusions: High tumour grade and larger size were shown to independently predict earlier breast cancer relapse. While LN involvement increases absolute recurrence risk, our study proposes that it does not influence timing of relapse. Use of these predictors will enable key risk periods for onset of relapse to be characterised according to tumour profile with more appropriate discharge to primary care providers for ongoing surveillance.