Molecular diagnosis of pancreatic cancer

Coordinator:		PD Dr. Malte Buchholz
Institution:		Philipps-Universität Marburg,Klinik für Innere Medizin SP Gastroenterologie
Homepage:		www.uni-marburg.de

To date, no single biomarker has been described which facilitates the accurate distinction between malignant and non-malignant processes in the pancreas. Efforts to improve the accuracy of pre-operative diagnostics by analyzing singular RNA, DNA or protein markers in pancreatic juice, brush cytologies or FNAB’s have regularly met with little success due to low sensitivity of the test, low specificity of the test, or both. Our own results as well as numerous reports in the literature demonstrate that the parallel analysis of different markers in a single step can significantly improve the accuracy of molecular diagnosis. As an example, we have demonstrated that querying a 169 gene signature using specialized diagnostic cDNA arrays is suitable to distinguish between adenocarcinoma of the pancreas and benign processes with >= 95 % accuracy.
In this subproject, we have made use of the central resources provided by TP1 to achieve two central goals: First, we have used a large number of “ideal” biopsy samples to validate methods and algorithms providing accurate distinction between 5 different forms of pancreatic tumors in a single diagnostic step (Gress et al. (2012), J. Mol. Med. 90: 457-464). Second, this diagnostic principle was transferred from the rather experimental cDNA array technology platform, which is not ideally suited for broad routine use, to the commercial TaqMan® low density array platform. Together with a dedicated analysis software, an integrated diagnostic tool is created that is straightforward to use, standardized across laboratories and suitable for use in routine clinical diagnostics. The development of a pancreatic cancer TaqMan® diagnostic array is thus a prime example of the IG’s fundamental goal of translating molecular knowledge into clinical applications with direct benefit for pancreatic cancer patients, and will indeed represent one of the first examples worldwide of a clinical tool developed on the basis of high-throughput genomic data.