Proteogenomic Metabolism Dependent Signature Identifies Worse Outcome in Breast And Other Cancers

Abstract

Breast cancer (BrCA) therapeutic selection routinely incorporates clinicopathologic information along with immunohistochemistry (IHC) for ER/PR/HER2/Ki-67. However, this is incomplete and has shortcomings that are seen in clinical outcome differences even within the same subtype. Herein, we analyzed the proteome of 116 HER2-negative primary BrCA samples and subsequently validated a 34-proteogenomic signature in 5,963 BrCA tumor samples from TCGA, METABRIC, and GSE96058 that demonstrated a metabolic enrichment signature impacting overall survival, progression free survival, and response to therapy. The 34-proteogenomic signature selected ER+ BrCA tumors for upstaging to a more triple negative pathophysiological phenotype, herein referred to as Luminal/TN-like (L/T), impacting likelihood for chemotherapy consideration and other therapeutic modalities rather than hormonal therapy alone. Further, analysis of 9,530 tumors across 33 types of cancers in TCGA demonstrated the 34 proteogenomic signature utility in the reclassification of other cancer types into different risk groups.