Abstract S1-06: Increased tumor-associated lymphocytes predict benefit from addition of carboplatin to neoadjuvant therapy for triple-negative and HER2-positive early breast cancer in the GeparSixto trial (GBG 66)

Abstract

Abstract Background: We have recently described a significantly increased pCR rate in triple-negative breast cancer with addition of carboplatin to a non-pegylated liposomal doxorubicin/taxane (MC) combination in the neoadjuvant Geparsixto study (von Minckwitz et al, ASCO, 2013). Here we report the results of prospective biomarker analyses performed in Geparsixto. Methods: Geparsixto investigates the effect of adding carboplatin to MC for the treatment of patients with HER2+ve and triple-negative (TN) primary BC. All HER2+ve patients received trastuzumab and lapatinib, all TN patients received bevacizumab. As part of the central pathology assessment, we prospectively evaluated tumor-associated lymphocytes (TILs) in 580 core biopsies. We used the parameters stromal lymphocytes (strLy) and lymphocyte-predominant breast cancer (LPBC, ≥60%TILs) as previously described (Denkert et al, JCO, 2010). In addition, Ki67 (n = 588) was measured by immunohistochemistry and p53 mutations in Exons 5-8 (n = 444) were evaluated by Sanger sequencing. Pathological complete remission (pCR) was defined as ypT0ypN0. Results: 24.7% of the 588 patients had a LPBC. These patients had an increased pCR rate of 60.3%, compared to 35.8% for the non-LPBC tumors (p<0.0005). A similar result was observed for strLy as a continuous parameter, with an odds ratio of 1.21 (95%CI 1.13-1.30), per 10% increase in strLy (p<0.0005). The effect of the lymphocytic infiltrate was different in both therapy arms: For patients treated with MC, the pCR rate was moderately increased in LPBC (46.6% vs. 33.5% in non-LPBC tumors, p = 0.050). In contrast, in patients who received additional carboplatin to MC, the pCR rate was increased to 75% for LPBC tumors, compared to 38.1% for non-LPBC tumors (p<0.0005). This result was observed in both therapy groups: In TNBC patients treated with MC+carboplatin, the pCR rates were 76.2% for LPBC, compared to 52.2% for non-LPBC (p = 0.01). In Her2+ve patients treated with MC+carboplatin, the pCR rates were 73.1% for LPBC and 22.9% for non-LPBC (p<0.0005). In the multivariate analysis, LPBC (p<0.0005) as well as strLy (p<0.0005) were a significant predictor of pCR after adjustment for tumor size, nodal status, therapy, age, hormone receptor and HER2 status. Ki67>20% was associated with an increased pCR rate of 46.6%, compared to 27% in tumors with lower proliferation (p<0.0005). For Ki67, the pCR rates were similar in both therapy arms. Mutations in exons 5-8 of the p53 gene were not associated with changes in pCR rate, neither in the total population nor in the two treatment arms. Conclusion: Our results show that the interaction with host immune response is relevant for response to chemotherapy, this effect is particularly strong with the addition of carboplatin. Tumor-associated lymphocytes as a continuous parameter as well as LPBC as a tumor subgroup are predictive for response to neoadjuvant chemotherapy. In particular with regard to the relevant toxicity of the MC+carboplatin combination, the integration of immunological biomarkers would be helpful to identify the patients with the highest benefit from the addition of carboplatin. Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr S1-06.