High amplified in breast cancer 1 is a significant predictor of improved response to adjuvant tamoxifen in premenopausal women.

Abstract

Abstract Abstract #6049 Background: The estrogen receptor (ER) coactivator amplified in breast cancer 1 (AIB1) has oncogene potentials and is believed to be of importance in breast cancer development and progression. Clinical trials implicate AIB1 to be a prognostic and a treatment predictive factor, although results have not been unanimous. This is to our knowledge the first report on AIB1's potential as a prognostic and/or a tamoxifen treatment predictive factor in a controlled randomised trial of tamoxifen vs no tamoxifen in premenopausal women.
 Patients and Methods: 349 premenopausal women with stage two breast cancer were randomised to tamoxifen for two years vs no tamoxifen independent of ER and progesterone receptor (PgR) status. Recurrence-free survival (RFS) was chosen as the primary outcome and the median follow-up for recurrence-free survivors was 14 years. Using a tissue microarray AIB1, ER and PgR were analysed with immunohistochemistry (IHC). Considering the proportion and intensity of stained nuclei, 56% of the tumours were considered as having high AIB1. Data on HER2 status and Nottingham histological grade (NHG) were available. Statistical methods used were X2-test, the Kaplan-Meier method, log rank test and Cox regression.
 Results: High AIB1 was found to correlate with other factors of worse prognosis (high HER2, high NHG and presence of lymph node metastases), as well as ER- and PgR-negativity. In the control arm, high AIB1 was a negative prognostic factor for RFS by univariate analysis (HR 1.6 p=0.02). In ER+ patients it remained a significant factor also in multivariate analyses including age, NHG, HER2, presence of lymph node metastases and tumour size (HR 1.8 p=0.03).
 Exploring the treatment predictive effect of AIB1, we found ER+ patients with high AIB1 to respond very well to tamoxifen treatment (HR 0.4 p=0.002), increasing RFS to the same levels as for systemically untreated patients with low AIB1. Statistical significance persisted also in a multivariate analysis (HR 0.3 p<0.001). Using a treatment interaction analysis, adjusted for other prognostic factors as above, we found that in ER+ patients the effect of tamoxifen was three times better in patients with high AIB1 than in those with low AIB1, a statistically significant difference ( p=0.007). ER+ patients with low AIB1 had a better RFS without treatment, but this was not further improved by tamoxifen (HR 0.9 p=0.8).
 Discussion: In this randomized trial of premenopausal breast cancer patients, we have shown AIB1 to be both a prognostic and a treatment predictive factor in the ER+ subgroup. Previous clinical studies regarding AIB1 have not shown unanimous results. Discrepancies maybe explained by differences in methodology and cut-off, that previous studies were not based on randomized trials, and/or that only premenopausal patients were included in the present study. Citation Information: Cancer Res 2009;69(2 Suppl):Abstract nr 6049.