GSE24279

• A solid process for diagnosis could have a substantial impact on the successful treatment of pancreatic cancer, for which currently mortality is nearly identical to incidence. 

• Variations in the abundance of all microRNA molecules from peripheral blood cells and pancreas tissues were analyzed on microarrays and in part validated by real-time PCR assays. 

• Utilizing the minimally invasive blood test, receiver operating characteristic (ROC) curves and the corresponding area under the curve (AUC) analysis demonstrated very high sensitivity and specificity of a distinction between healthy people and patients with either cancer or chronic pancreatitis; respective AUC values of 0.973 and 0.950 were obtained. Confirmative and partly even more discriminative diagnosis could be performed on tissue samples with AUC values of 1.0 and 0.937, respectively. 

• In addition, discrimination between cancer and chronic pancreatitis was achieved (AUC = 0.875). 

• A classification model has been built using Trainset. The selected probes were:
  1. hsa-mir-181a
  2. hsa-mir-155

• The model has been tested using Testset.