Measurement of baseline serum SDF-1 levels as a predictive biomarker for outcomes in the NCIC CTG MA.14 trial of octreotide, a somatostatin analogue in postmenopausal breast cancer.

Abstract

577 Background: Stromal cell derived-factor 1 (SDF-1) is a chemokine implicated in breast tumor growth and metastasis. Blood SDF-1 levels have a significant prognostic value in primary breast cancers. We have reported that 4 month serum SDF-1 levels had a significant EFS predictive value in the adjuvant breast cancer population of MA.14 (ASCO 2010). Here, we sought to confirm the association with baseline assessments of SDF-1. Methods: NCIC CTG MA.14 is a randomized phase III trial in postmenopausal women of tam 20 mg daily for 5 yr +/- octreotide LAR for 2 yr. Stratification factors were adjuvant chemo, axillary nodal status, and ER status. The primary endpoint was event-free survival (EFS) defined as time from randomization to recurrence, second malignancy, or death due to any cause. Serum was collected at various timepoints throughout the trial. Baseline SDF-1 serum levels were measured with an ELISA kit (R&D). To assess the effect of SDF-1 on EFS in the context of factors considered in the main trial, we used step-wise forward multivariate adjusted Cox modeling, adjusting by the stratification factors, and adding a factor if the likelihood ratio criterion was significant, p ≤ 0.05. Events for the analysis were those reported at ASCO 2008. Results: 667 women were accrued (333 to tam, 334 to tam + oct), and followed for a median 7.9 years. We successfully measured the baseline levels of SDF-1 in 469 (70%) patients, 241 on tamoxifen arm, of whom 80 (33%) experienced an EFS event; and for 228 on tamoxifen and octreotide LAR arm, of whom 70 (31%) had an EFS event.Consistent with our previous findings, we found that in a step-wise multivariate analysis, higher log (SDF-1) (p < 0.0001) was significantly associated with worse EFS. Also, the interaction of treatment and SDF-1 split at median was significantly (p < 0.0001) associated with EFS with inclusion of the significant main trial factors. Conclusions: We have confirmed that baseline SDF-1 levels have a predictive value in the breast cancer population of MA.14 treated with octreotide LAR. Further validation of SDF-1 and investigation of a possible pharmacogenetic effect with octreotide therapy is warranted.