Comparative proteomic profiling of the serum differentiates pancreatic cancer from chronic pancreatitis

Abstract

Finland ranks sixth among the countries having highest incidence rate of pancreatic cancer with mortality roughly equaling incidence. The average age of diagnosis for pancreatic cancer is 69 years in Nordic males, whereas the average age of diagnosis of chronic pancreatitis is 40–50 years, however, many cases overlap in age. By radiology, the evaluation of a pancreatic mass, that is, the differential diagnosis between chronic pancreatitis and pancreatic cancer is often difficult. Preoperative needle biopsies are difficult to obtain and are demanding to interpret. New blood based biomarkers are needed. The accuracy of the only established biomarker for pancreatic cancer, CA 19-9 is rather poor in differentiating between benign and malignant mass of the pancreas. In this study, we have performed mass spectrometry analysis (High Definition MS^E) of serum samples from patients with chronic pancreatitis (13) and pancreatic cancer (22). We have quantified 291 proteins and performed detailed statistical analysis such as principal component analysis, orthogonal partial least square discriminant analysis and receiver operating curve analysis. The proteomic signature of chronic pancreatitis versus pancreatic cancer samples was able to separate the two groups by multiple statistical techniques. Some of the enriched pathways in the proteomic dataset were LXR/RXR activation, complement and coagulation systems and inflammatory response. We propose that multiple high-confidence biomarker candidates in our pilot study including Inter-alpha-trypsin inhibitor heavy chain H2 (Area under the curve, AUC: 0.947), protein AMBP (AUC: 0.951) and prothrombin (AUC: 0.917), which should be further evaluated in larger patient series as potential new biomarkers for differential diagnosis.

Serum proteomics has the potential to find the candidate biomarkers for various types of cancers enabling differential diagnosis. We have performed HDMSE of serum samples from chronic pancreatitis and pancreatic cancer patients and quantified hundreds of proteins. Exhaustive statistical analyses such as OPLS-DA, PCA, ROC curve analyses and multiple pathway analyses were performed which provided dysregulated pathways in these diseases and high-confidence potential biomarker candidates were proposed.

http://ift.tt/2s02JQ7