MicroRNA expression patterns to differentiate pancreatic adenocarcinoma from normal pancreas and chronic pancreatitis.

ABSTRACT

CONTEXT While global microRNA (miRNA) expression patterns of many embryologic, physiologic, and oncogenic processes have been described, description of the role of miRNAs in ductal adenocarcinoma of the pancreas is lacking. OBJECTIVE: To define the expression pattern of miRNAs in pancreatic cancer and compare it with those of normal pancreas and chronic pancreatitis. DESIGN AND SETTING: Specimens were obtained at a National Cancer Institute-designated comprehensive cancer center from patients with ductal adenocarcinoma of the pancreas (n = 65) or chronic pancreatitis (n = 42) (January 2000-December 2005). All patients underwent curative pancreatectomy; those with pancreatic cancer were chemotherapy-naive. RNA harvested from resected pancreatic cancers and matched benign adjacent pancreatic tissue as well as from chronic pancreatitis specimens was hybridized to miRNA microarrays. MAIN OUTCOME MEASURES: Identification of differentially expressed miRNAs that could differentiate pancreatic cancer from normal pancreas, chronic pancreatitis, or both, as well as a pattern of miRNA expression predictive of long-term (>24 months) survival. Significance of Analysis of Microarrays and Prediction of Analysis of Microarrays were undertaken to identify miRNAs predictive of tissue type and prognosis. P values were calculated by t test, adjusted for multiple testing. Kaplan-Meier survival curves were constructed using mean miRNA expression (high vs low) as threshold and compared by log-rank analysis. RESULTS: Twenty-one miRNAs with increased expression and 4 with decreased expression were identified that correctly differentiated pancreatic cancer from benign pancreatic tissue in 90% of samples by cross validation. Fifteen overexpressed and 8 underexpressed miRNAs differentiated pancreatic cancer from chronic pancreatitis with 93% accuracy. A subgroup of 6 miRNAs was able to distinguish long-term survivors with node-positive disease from those dying within 24 months. Finally, high expression of miR-196a-2 was found to predict poor survival (median, 14.3 months [95% confidence interval, 12.4-16.2] vs 26.5 months [95% confidence interval, 23.4-29.6]; P = .009). CONCLUSIONS: Pancreatic cancer may have a distinct miRNA expression pattern that may differentiate it from normal pancreas and chronic pancreatitis. miRNA expression patterns may be able to distinguish between long- and short-term survivors, but these findings need to be validated in other study populations.