Purpose: Metastatic breast cancer (MBC) is not curable and there is a growing interest in personalized therapy options. Here we report molecular profiling of MBC focusing on molecular evolution in actionable alterations. Experimental Design: Sixty-two patients with MBC were included. An analysis of DNA, RNA, and functional proteomics was done, and matched primary and metastatic tumors were compared when feasible. Results: Targeted exome sequencing of 41 tumors identified common alterations in TP53 (21; 51%) and PIK3CA (20; 49%), as well as alterations in several emerging biomarkers such as NF1 mutations/deletions (6; 15%), PTEN mutations (4; 10%), and ARID1A mutations/deletions (6; 15%). Among 27 hormone receptor–positive patients, we identified MDM2 amplifications (3; 11%), FGFR1 amplifications (5; 19%), ATM mutations (2; 7%), and ESR1 mutations (4; 15%). In 10 patients with matched primary and metastatic tumors that underwent targeted exome sequencing, discordances in actionable alterations were common, including NF1 loss in 3 patients, loss of PIK3CA mutation in 1 patient, and acquired ESR1 mutations in 3 patients. RNA sequencing in matched samples confirmed loss of NF1 expression with genomic NF1 loss. Among 33 patients with matched primary and metastatic samples that underwent RNA profiling, 14 actionable genes were differentially expressed, including antibody–drug conjugate targets LIV-1 and B7-H3. Conclusions: Molecular profiling in MBC reveals multiple common as well as less frequent but potentially actionable alterations. Genomic and transcriptional profiling demonstrates intertumoral heterogeneity and potential evolution of actionable targets with tumor progression. Further work is needed to optimize testing and integrated analysis for treatment selection.
1005 Background: By gene profiling, Lehmann et al. (J Clin Invest 121:2750-2767, 2011) reported that triple-negative breast cancer (TNBC) can be classified into 6 clusters—basal-like 1 (BL1), basal-like 2 (BL2), immunomodulatory (IM), mesenchymal (M), mesenchymal stem-like (MSL), and luminal androgen receptor (LAR)—plus an unstable (UNS) cluster. While it is clear that patients with TNBC differently respond to chemotherapy, the clinical relevance of these molecular TNBC subtypes is unknown. Methods: We qualitatively reproduced the Lehmann et al. experiments using Affymetrix CEL files from the public datasets. We identified 130 TNBC gene expression microarrays obtained from 03/00 to 03/10. All patients had received neoadjuvant chemotherapy containing sequential taxane and anthracycline-based regimens and had evaluable pathological tumor response data. Median follow-up was 68.1 months. (5.1-147.5). We classified TNBC samples using Lehmann’s gene signatures, then performed Fisher's exact test to correlate TNBC subtype and pCR status. To assess the independent utility of TNBC subtype for predicting pCR status, we fit a logistic regression model to our data and used age, clinical stage, treatment regimens, and nuclear grade as potential explanatory factors. We also performed comparison of the subtypes with the PAM50 intrinsic subtypes and RCB index. Results: The BL1 subtype had the highest pCR rate (52%); BL2 and LAR the lowest (0% and 10%, respectively). TNBC subtype and pCR status were significantly associated (p = 0.044). TNBC subtype was an independent predictor of pCR status (p = 0.022) by a likelihood ratio test.The Lehmann’s subtype classification better predicted pCR status than did the PAM50 intrinsic subtypes (basal-like vs. non basal-like). Conclusions: Dividing TNBC into 7 subtypes predicts high vs. low pCR rate. The 7-subtype classification may lead to innovative personalized medicine strategies for patients with TNBC. There is a need for prospective validation of the hypothesis that pCR rates associated with the seven TNBC subtypes will predict long-term patient outcome.