Abstract 1385: Molecular changes in breast tumors following bevacizumab-based treatment: Final analysis of a randomized neoadjuvant study of bevacizumab or placebo, followed by chemotherapy with or without bevacizumab, in patients with stage II or III breast cancer

Proceedings: AACR 103rd Annual Meeting 2012‐‐ Mar 31‐Apr 4, 2012; Chicago, IL Background: Bevacizumab (bev) has been widely studied in breast cancer (BC), yet no randomized trial in BC has reported in vivo molecular effects of bev on human tumor tissue. We conducted a trial to evaluate the safety, clinical & molecular effects of neoadjuvant chemotherapy plus bev for locally advanced BC. Methods: 90 pts were randomized (2:1:2:1) to 1 of 4 arms: Arm A: TAC (docetaxel, T: 75 mg/m2, doxorubicin, A: 50 mg/m2, cyclophosphamide, C: 500 mg/m2) + low dose bev (7.5 mg/kg); Arm B: TAC + low dose placebo (P); Arm C: TAC + standard dose bev (15 mg/kg); Arm D: TAC + Std-P. A run-in cycle of bev or P was followed by 6 cycles of TAC plus P or bev. Tumor biopsies pre- and 7-10 days post-run-in with bev or P were taken. Unblinding occurred post surgery (Sx): Arms A/C received maintenance bev to complete 52 wks, Arms B/D received no further P. Eligible patients were females with >3 cm, HER2(-) BC. Endpoints included safety and pathologic complete response (pCR) in breast & lymph nodes. To assess the effects of VEGF pathway inhibition on tumor vasculature, PCR analysis using the fluidigm array platform was performed on RNA from the pre- and post-run in samples to evaluate expression of 67 genes known to play a defined role in VEGF signaling. Results: 28 pts were assigned to Arm A, 30 to Arm C, and 32 to Arms B/D. 12 pts came off Tx before surgery (Arm A:2, Arm C:6, Arms B/D:4) and 78 received all Tx, underwent Sx and are evaluable for pCR. Two pts (6%) in Arms B/D, 5 pts (18%) in Arm A and 10 pts (33%) in Arm C had wound healing complications. 17% of pts in Arm C had Gr 3 (N=4) or Gr 4 (N=1) heart failure, none in Arms A/B/D. The pCR rate in evaluable patients was 18% (14/78): 5 (19%) Arm A, 3 (13%) Arm C, 6 (21%) Arms B/D. 45 samples (20 from Arms B/D and 25 from bev arms) were included in the pair-wise analysis to identify 6 genes that were differentially expressed. Bev resulted in a decrease in expression of Dll4, Cox2, Fibronectin (FN_EIIIB), angiopoietin 2 (Angpt2) and ESM1. In addition, upregulation of the cytokine, stromal derived growth factor (SDF1) was observed. Notably, genes such as CD31 or VE-cadherin were not appreciably differentially expressed. Of the genes that were downregulated, Dll4 and Angpt2 represent genes enriched in endothelial tip cells, which guide the migration of newly formed blood vessels. The decrease in these genes likely represents the effect of bev on reducing immature, growing vasculature in the tumor. Conclusions: This trial enabled the clinical and molecular evaluation of breast cancer tumor tissue pre and post-bev. Clinically, bev was associated with more wound healing and heart failure events and similar pCR rates compared to P. Tumor expression analysis for genes in the angiogenesis pathway supports the preclinical hypothesis that bev may primarily target immature tumor vasculature. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 1385. doi:1538-7445.AM2012-1385