Everolimus is an agent frequently associated with specific toxicities. Predictive markers of efficacy are needed to help define which patients could benefit from it. The goal of this exploratory study was to identify potential predictive biomarkers in the mammalian target of rapamycin (mTOR) complex 1 (mTORC1) activation pathway using primary tumor samples collected during the phase II tamoxifen plus everolimus (TAMRAD) trial.Tumor tissues were collected retrospectively from the TAMRAD trial. Immunohistochemistry was carried out using specific antibodies directed toward proteins that result in mTORC1 activation [canonical phosphatidylinositol 3-kinase (PI3K)/protein kinase B (Akt)/mTOR or alternative pathways]. DNA was extracted from the tumor tissue; mutation screening in the PIK3CA gene (exons 9 and 20) and the KRAS gene (exons 2 and 3) was first carried out using Sanger direct sequencing, and then completed by next-generation sequencing for PIK3CA. An exploratory analysis of everolimus efficacy in terms of a time-to-progression (TTP) increase was carried out in each biomarker subgroup (high versus low expression referring to the median percentage of marked cells).A total of 55 primary tumor samples from the TAMRAD trial-25 from the tamoxifen-alone group and 30 from the tamoxifen/everolimus group-were evaluated for biomarkers. The subgroups most likely to have an improvement in TTP with tamoxifen/everolimus therapy, compared with tamoxifen alone, were patients with high p4EBP1, low 4EBP1, low liver kinase B1, low pAkt, and low PI3K. Among the 45 samples screened for mutation status, nine samples (20%; 95% CI 9.6-34.6) had a PIK3CA mutation. KRAS mutation was observed in one patient.A positive correlation between late effectors of mTORC1 activation, a positive correlation between Akt-independent mTORC1 activation, and an inverse correlation between canonical PI3K/Akt/mTOR pathway and everolimus efficacy were observed in this exploratory analysis. However, these correlations need to be validated in larger studies before applying the findings to routine clinical practice.

Abstract Background: TNBC is a heterogenous group of tumors for some of which the Epithelial Growth Factor Receptor pathway (EGFR) may play an important role. We evaluated the efficacy and toxicity of an anti-EGFR antibody (cetuximab) combined with docetaxel, which were given to the TNBC patients (pts) in the neoadjuvant setting. A biomarker analysis accompanied this trial, aiming to identify predictors of response. Methods: 35 patients with stage II-IIIA TN breast disease were prospectively included in this multicentre pilot study. Systemic therapy (ST) consisted of 6 cycles of docetaxel (100 mg/m2) each 3 weeks, in combination with weekly cetuximab (first dose 400mg/m2 then 250 mg/m2/week) for 6 cycles. All patients underwent surgery at completion of ST. Patient characteristics : mean age 48 [28-67]; TNM: T1: 3%, T2: 73%, T3: 24%; N0: 61%, N1-N2: 39%; mean tumor size 40 mm [15-100]; SBR: grade III: 73%, grade II: 27%. The median number of cycles was for docetaxel 6 [1-6] and for cetuximab 15 [1-18]. Pathological complete response (pCR) rate was 24% according to the Chevallier and Sataloff classifications; 28% if we consider breast pCR only. Overall clinical response rate was 57% (22% CR). Conservative surgery was performed in 75% of cases. The main side effect was skin toxicity: grade II: 39%, grade III: 36%, grade IV: 3%. Other side effects were: neutropenia grade IV: 12.7%, febrile neutropenia: 1.3%, hand-foot syndrome grade III: 3%, grade II: 3%, ungueal toxicity grade III: 3%, grade II: 33%. Paraffin-embedded and frozen samples were systematically collected before and after ST for biomarker studies. Germinal BRCA1 mutations and EGFR, KRAS, BRAF and PI3KCA somatic mutations were analyzed by NGS. EGFR, MET, cytokeratins 5/6 and 8/18, PTEN, P-cadherin, ALDH1, Ki67, p53 and the number of FOXP3+ or CD8+ tumor-infiltrating lymphocytes (TIL) were evaluated by immunohistochemistry. Results: The biomarker analysis was interpretable on the samples from 21 pts (3 pCR and 18 non-pCR). We applied the ROC curve to identify the best cut-off value for Ki67, EGFR, MET, cytokeratin 5/6 and 8/18, p53, ALDH1, PTEN, P-cadherin and the FOXP3+ or CD8+ TIL counts. None of these biomarkers was predictive of pCR except for the CD8+/FOXP3+ TIL count ratio. pCR rate was higher in the pts with the ratio equal or higher than 2.75 than in the others (43% versus 0%, p = 0.047). BRCA1 mutations were detected in 16% of pts. PI3K and EGFR somatic mutations were observed in 1 and 3 patients, respectively. The presence of the mutations was not predictive of pCR. Conclusion: Similarly to the previously reported trial by our group (SABCS 2012, abstract 1081), the immune component of the tumor microenvironment plays a very important role in the TNBC response to cytotoxic therapy. Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr P1-08-34.