Integrative Survival-Based Molecular Profiling of Human Pancreatic Cancer [mRNA]

To perform an integrative profile of human pancreatic cancer (PDAC) to identify prognosis-significant genes and their related pathways. A concordant survival-based whole genome in silico array analysis of DNA copy number, and mRNA & micro RNA (miRNA) expression in 25 early stage PDAC was performed. A novel composite score simultaneously integrated gene expression with regulatory mechanisms to identify the signature genes with the most levels of prognosis-significant evidence. The predominant signaling pathways were determined via a pathway-based approach. Independent patient cohorts (n= 150 and 42) were then used as in vitro validation of the array findings. We find that EGFR, SRC signaling, and PI3K/AKT pathway activation are strongly linked to clinical disease progression. Furthermore, we identify two discrete subsets of pancreatic tumors characterized by either SRC or PI3K/AKT signaling that may dictate variable responses to targeted therapy. 42 human PDAC tumors and 7 non-malignant pancreas samples snap-frozen at the time of surgery were chosen. Representative H&E sections of each were evaluated by a practicing gastrointestinal pathologist to confirm diagnosis and determine relative percentage of malignant cells. Samples with tumor cell content >30% were chosen for final multi-platform analysis (N=25). Human pancreatic samples were profiled on Affymetrix HGU133 Plus 2 arrays per manufacturer's instructions.