Background: The I-SPY trial is a multi-institutional study of locally advanced breast cancers in which serial biopsies of breast tumors are performed at multiple timepoints, allowing comparison of tumor biomarker status before, during, and after therapy. Tumor HER2 status has an important role in predicting response to chemotherapy and targeted therapy in breast cancer. The aim of this study was to determine if the administration of neoadjuvant chemotherapy was associated with changes in tumor HER2 protein expression.

Methods: Eligible patients had biopsy-confirmed primary breast cancer measuring at least 3 cm. Required initial treatment was an anthracycline based neoadjuvant chemotherapy regimen: AC followed by T (91%), AC without T (6%), and chemotherapy beyond AC +/- T 2%. Serial core biopsies were performed at baseline (T1), after one cycle of AC (T2), intra-regimen (T3), and pre-surgery (T4). Immunohistochemistry (IHC) for HER2 was performed centrally on core samples at T1 and T2 and on post-surgery (PS) excision/mastectomy specimens (for patients with residual tumor). HER2 results were classified as negative, indeterminate, or overexpressed using conventional scoring criteria. Tumor HER2 status was compared between timepoints T1 and T2, T2 and PS, and T1 and PS. Patients receiving Herceptin therapy were excluded from this analysis. Additional molecular analysis is being performed and the IHC results will be compared with gene expression and phospho-HER2 results.

Results: 221 patients were enrolled and completed therapy. The median age was 49 years, with a range of 27-69. Median tumor size was 6 cm; 58% had clinically positive nodes. HER2 results were obtained centrally at baseline on 178 (80%) patients. 41 of 178 (23%) tumors were positive for HER2 overexpression. The overall pCR rate in HER2 positive patients was 52% compared to 19% for HER2 negative patients (p<0.001). Patient HER2 results were compared at T1 and T2 (n=154), T2 and PS (n=81), and T1 and PS (n=89). Nine of the 32 (28%) HER2 positive patients with multiple timepoint HER2 results were excluded due to receiving Herceptin therapy. The tumor HER2 status between timepoints is summarized in the table below.

Conclusions: Tumor HER2 protein expression levels appear stable in patients undergoing neoadjuvant chemotherapy with anthracycline- and taxane-based regimens. Rare tumors (2-3%) showed a change from a HER2 positive result to a HER2 negative result during chemotherapy; these findings may be related to tumor heterogeneity rather than a true treatment-induced change in HER2 status.